{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.893964110929853,
  "eval_steps": 50,
  "global_step": 18000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0027188689505165853,
      "grad_norm": 8.0,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.7742,
      "step": 10
    },
    {
      "epoch": 0.005437737901033171,
      "grad_norm": 8.8125,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.7824,
      "step": 20
    },
    {
      "epoch": 0.008156606851549755,
      "grad_norm": 4.15625,
      "learning_rate": 1.5e-06,
      "loss": 0.7676,
      "step": 30
    },
    {
      "epoch": 0.010875475802066341,
      "grad_norm": 10.125,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.7742,
      "step": 40
    },
    {
      "epoch": 0.013594344752582926,
      "grad_norm": 4.8125,
      "learning_rate": 2.5e-06,
      "loss": 0.7496,
      "step": 50
    },
    {
      "epoch": 0.013594344752582926,
      "eval_loss": 0.7354375123977661,
      "eval_runtime": 136.6716,
      "eval_samples_per_second": 3.658,
      "eval_steps_per_second": 0.117,
      "step": 50
    },
    {
      "epoch": 0.01631321370309951,
      "grad_norm": 2.828125,
      "learning_rate": 3e-06,
      "loss": 0.7281,
      "step": 60
    },
    {
      "epoch": 0.019032082653616094,
      "grad_norm": 2.078125,
      "learning_rate": 3.5000000000000004e-06,
      "loss": 0.7305,
      "step": 70
    },
    {
      "epoch": 0.021750951604132682,
      "grad_norm": 1.8046875,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.7172,
      "step": 80
    },
    {
      "epoch": 0.024469820554649267,
      "grad_norm": 1.484375,
      "learning_rate": 4.5e-06,
      "loss": 0.7027,
      "step": 90
    },
    {
      "epoch": 0.027188689505165852,
      "grad_norm": 1.2734375,
      "learning_rate": 5e-06,
      "loss": 0.7094,
      "step": 100
    },
    {
      "epoch": 0.027188689505165852,
      "eval_loss": 0.702218770980835,
      "eval_runtime": 112.7091,
      "eval_samples_per_second": 4.436,
      "eval_steps_per_second": 0.142,
      "step": 100
    },
    {
      "epoch": 0.029907558455682437,
      "grad_norm": 1.0625,
      "learning_rate": 5.500000000000001e-06,
      "loss": 0.6984,
      "step": 110
    },
    {
      "epoch": 0.03262642740619902,
      "grad_norm": 1.140625,
      "learning_rate": 6e-06,
      "loss": 0.6969,
      "step": 120
    },
    {
      "epoch": 0.03534529635671561,
      "grad_norm": 1.1484375,
      "learning_rate": 6.5000000000000004e-06,
      "loss": 0.6957,
      "step": 130
    },
    {
      "epoch": 0.03806416530723219,
      "grad_norm": 1.125,
      "learning_rate": 7.000000000000001e-06,
      "loss": 0.682,
      "step": 140
    },
    {
      "epoch": 0.040783034257748776,
      "grad_norm": 0.98046875,
      "learning_rate": 7.5e-06,
      "loss": 0.6832,
      "step": 150
    },
    {
      "epoch": 0.040783034257748776,
      "eval_loss": 0.6807812452316284,
      "eval_runtime": 109.6347,
      "eval_samples_per_second": 4.561,
      "eval_steps_per_second": 0.146,
      "step": 150
    },
    {
      "epoch": 0.043501903208265365,
      "grad_norm": 1.1953125,
      "learning_rate": 8.000000000000001e-06,
      "loss": 0.6797,
      "step": 160
    },
    {
      "epoch": 0.046220772158781946,
      "grad_norm": 0.80078125,
      "learning_rate": 8.500000000000002e-06,
      "loss": 0.6699,
      "step": 170
    },
    {
      "epoch": 0.048939641109298535,
      "grad_norm": 1.0859375,
      "learning_rate": 9e-06,
      "loss": 0.6777,
      "step": 180
    },
    {
      "epoch": 0.051658510059815116,
      "grad_norm": 1.1875,
      "learning_rate": 9.5e-06,
      "loss": 0.6602,
      "step": 190
    },
    {
      "epoch": 0.054377379010331704,
      "grad_norm": 1.5859375,
      "learning_rate": 1e-05,
      "loss": 0.6582,
      "step": 200
    },
    {
      "epoch": 0.054377379010331704,
      "eval_loss": 0.6568124890327454,
      "eval_runtime": 107.0118,
      "eval_samples_per_second": 4.672,
      "eval_steps_per_second": 0.15,
      "step": 200
    },
    {
      "epoch": 0.057096247960848286,
      "grad_norm": 1.5,
      "learning_rate": 1.05e-05,
      "loss": 0.6496,
      "step": 210
    },
    {
      "epoch": 0.059815116911364874,
      "grad_norm": 1.4453125,
      "learning_rate": 1.1000000000000001e-05,
      "loss": 0.6426,
      "step": 220
    },
    {
      "epoch": 0.06253398586188146,
      "grad_norm": 1.59375,
      "learning_rate": 1.1500000000000002e-05,
      "loss": 0.6402,
      "step": 230
    },
    {
      "epoch": 0.06525285481239804,
      "grad_norm": 2.328125,
      "learning_rate": 1.2e-05,
      "loss": 0.6234,
      "step": 240
    },
    {
      "epoch": 0.06797172376291463,
      "grad_norm": 2.671875,
      "learning_rate": 1.25e-05,
      "loss": 0.6121,
      "step": 250
    },
    {
      "epoch": 0.06797172376291463,
      "eval_loss": 0.6246562600135803,
      "eval_runtime": 106.2642,
      "eval_samples_per_second": 4.705,
      "eval_steps_per_second": 0.151,
      "step": 250
    },
    {
      "epoch": 0.07069059271343121,
      "grad_norm": 3.1875,
      "learning_rate": 1.3000000000000001e-05,
      "loss": 0.6031,
      "step": 260
    },
    {
      "epoch": 0.0734094616639478,
      "grad_norm": 2.46875,
      "learning_rate": 1.3500000000000001e-05,
      "loss": 0.5941,
      "step": 270
    },
    {
      "epoch": 0.07612833061446438,
      "grad_norm": 2.328125,
      "learning_rate": 1.4000000000000001e-05,
      "loss": 0.5918,
      "step": 280
    },
    {
      "epoch": 0.07884719956498097,
      "grad_norm": 3.21875,
      "learning_rate": 1.45e-05,
      "loss": 0.5914,
      "step": 290
    },
    {
      "epoch": 0.08156606851549755,
      "grad_norm": 4.3125,
      "learning_rate": 1.5e-05,
      "loss": 0.5895,
      "step": 300
    },
    {
      "epoch": 0.08156606851549755,
      "eval_loss": 0.6041250228881836,
      "eval_runtime": 104.0917,
      "eval_samples_per_second": 4.803,
      "eval_steps_per_second": 0.154,
      "step": 300
    },
    {
      "epoch": 0.08428493746601413,
      "grad_norm": 3.125,
      "learning_rate": 1.55e-05,
      "loss": 0.5867,
      "step": 310
    },
    {
      "epoch": 0.08700380641653073,
      "grad_norm": 3.0625,
      "learning_rate": 1.6000000000000003e-05,
      "loss": 0.5408,
      "step": 320
    },
    {
      "epoch": 0.08972267536704731,
      "grad_norm": 2.953125,
      "learning_rate": 1.65e-05,
      "loss": 0.5598,
      "step": 330
    },
    {
      "epoch": 0.09244154431756389,
      "grad_norm": 3.0,
      "learning_rate": 1.7000000000000003e-05,
      "loss": 0.5318,
      "step": 340
    },
    {
      "epoch": 0.09516041326808047,
      "grad_norm": 3.0625,
      "learning_rate": 1.75e-05,
      "loss": 0.5676,
      "step": 350
    },
    {
      "epoch": 0.09516041326808047,
      "eval_loss": 0.559499979019165,
      "eval_runtime": 132.2047,
      "eval_samples_per_second": 3.782,
      "eval_steps_per_second": 0.121,
      "step": 350
    },
    {
      "epoch": 0.09787928221859707,
      "grad_norm": 3.453125,
      "learning_rate": 1.8e-05,
      "loss": 0.5496,
      "step": 360
    },
    {
      "epoch": 0.10059815116911365,
      "grad_norm": 6.4375,
      "learning_rate": 1.85e-05,
      "loss": 0.5041,
      "step": 370
    },
    {
      "epoch": 0.10331702011963023,
      "grad_norm": 2.53125,
      "learning_rate": 1.9e-05,
      "loss": 0.5312,
      "step": 380
    },
    {
      "epoch": 0.10603588907014681,
      "grad_norm": 5.6875,
      "learning_rate": 1.9500000000000003e-05,
      "loss": 0.5326,
      "step": 390
    },
    {
      "epoch": 0.10875475802066341,
      "grad_norm": 4.28125,
      "learning_rate": 2e-05,
      "loss": 0.5152,
      "step": 400
    },
    {
      "epoch": 0.10875475802066341,
      "eval_loss": 0.523687481880188,
      "eval_runtime": 111.1715,
      "eval_samples_per_second": 4.498,
      "eval_steps_per_second": 0.144,
      "step": 400
    },
    {
      "epoch": 0.11147362697117999,
      "grad_norm": 4.4375,
      "learning_rate": 2.05e-05,
      "loss": 0.5227,
      "step": 410
    },
    {
      "epoch": 0.11419249592169657,
      "grad_norm": 4.25,
      "learning_rate": 2.1e-05,
      "loss": 0.5379,
      "step": 420
    },
    {
      "epoch": 0.11691136487221315,
      "grad_norm": 3.390625,
      "learning_rate": 2.15e-05,
      "loss": 0.5033,
      "step": 430
    },
    {
      "epoch": 0.11963023382272975,
      "grad_norm": 5.8125,
      "learning_rate": 2.2000000000000003e-05,
      "loss": 0.5311,
      "step": 440
    },
    {
      "epoch": 0.12234910277324633,
      "grad_norm": 3.203125,
      "learning_rate": 2.25e-05,
      "loss": 0.5143,
      "step": 450
    },
    {
      "epoch": 0.12234910277324633,
      "eval_loss": 0.484187513589859,
      "eval_runtime": 109.6169,
      "eval_samples_per_second": 4.561,
      "eval_steps_per_second": 0.146,
      "step": 450
    },
    {
      "epoch": 0.1250679717237629,
      "grad_norm": 4.21875,
      "learning_rate": 2.3000000000000003e-05,
      "loss": 0.4582,
      "step": 460
    },
    {
      "epoch": 0.1277868406742795,
      "grad_norm": 4.15625,
      "learning_rate": 2.35e-05,
      "loss": 0.4752,
      "step": 470
    },
    {
      "epoch": 0.13050570962479607,
      "grad_norm": 3.515625,
      "learning_rate": 2.4e-05,
      "loss": 0.4602,
      "step": 480
    },
    {
      "epoch": 0.13322457857531267,
      "grad_norm": 4.5625,
      "learning_rate": 2.45e-05,
      "loss": 0.4406,
      "step": 490
    },
    {
      "epoch": 0.13594344752582926,
      "grad_norm": 2.9375,
      "learning_rate": 2.5e-05,
      "loss": 0.4465,
      "step": 500
    },
    {
      "epoch": 0.13594344752582926,
      "eval_loss": 0.4515624940395355,
      "eval_runtime": 106.1096,
      "eval_samples_per_second": 4.712,
      "eval_steps_per_second": 0.151,
      "step": 500
    },
    {
      "epoch": 0.13866231647634583,
      "grad_norm": 6.28125,
      "learning_rate": 2.5500000000000003e-05,
      "loss": 0.4127,
      "step": 510
    },
    {
      "epoch": 0.14138118542686243,
      "grad_norm": 3.453125,
      "learning_rate": 2.6000000000000002e-05,
      "loss": 0.4463,
      "step": 520
    },
    {
      "epoch": 0.14410005437737902,
      "grad_norm": 3.5,
      "learning_rate": 2.6500000000000004e-05,
      "loss": 0.4545,
      "step": 530
    },
    {
      "epoch": 0.1468189233278956,
      "grad_norm": 2.625,
      "learning_rate": 2.7000000000000002e-05,
      "loss": 0.4238,
      "step": 540
    },
    {
      "epoch": 0.14953779227841218,
      "grad_norm": 4.46875,
      "learning_rate": 2.7500000000000004e-05,
      "loss": 0.4527,
      "step": 550
    },
    {
      "epoch": 0.14953779227841218,
      "eval_loss": 0.4073750078678131,
      "eval_runtime": 105.0038,
      "eval_samples_per_second": 4.762,
      "eval_steps_per_second": 0.152,
      "step": 550
    },
    {
      "epoch": 0.15225666122892875,
      "grad_norm": 4.53125,
      "learning_rate": 2.8000000000000003e-05,
      "loss": 0.3699,
      "step": 560
    },
    {
      "epoch": 0.15497553017944535,
      "grad_norm": 4.3125,
      "learning_rate": 2.8499999999999998e-05,
      "loss": 0.3875,
      "step": 570
    },
    {
      "epoch": 0.15769439912996194,
      "grad_norm": 3.515625,
      "learning_rate": 2.9e-05,
      "loss": 0.3965,
      "step": 580
    },
    {
      "epoch": 0.1604132680804785,
      "grad_norm": 6.96875,
      "learning_rate": 2.95e-05,
      "loss": 0.3881,
      "step": 590
    },
    {
      "epoch": 0.1631321370309951,
      "grad_norm": 4.6875,
      "learning_rate": 3e-05,
      "loss": 0.3818,
      "step": 600
    },
    {
      "epoch": 0.1631321370309951,
      "eval_loss": 0.38462498784065247,
      "eval_runtime": 102.273,
      "eval_samples_per_second": 4.889,
      "eval_steps_per_second": 0.156,
      "step": 600
    },
    {
      "epoch": 0.1658510059815117,
      "grad_norm": 3.40625,
      "learning_rate": 3.05e-05,
      "loss": 0.359,
      "step": 610
    },
    {
      "epoch": 0.16856987493202827,
      "grad_norm": 4.375,
      "learning_rate": 3.1e-05,
      "loss": 0.3854,
      "step": 620
    },
    {
      "epoch": 0.17128874388254486,
      "grad_norm": 3.515625,
      "learning_rate": 3.15e-05,
      "loss": 0.3914,
      "step": 630
    },
    {
      "epoch": 0.17400761283306146,
      "grad_norm": 3.953125,
      "learning_rate": 3.2000000000000005e-05,
      "loss": 0.3799,
      "step": 640
    },
    {
      "epoch": 0.17672648178357803,
      "grad_norm": 4.28125,
      "learning_rate": 3.2500000000000004e-05,
      "loss": 0.3771,
      "step": 650
    },
    {
      "epoch": 0.17672648178357803,
      "eval_loss": 0.35850000381469727,
      "eval_runtime": 103.4282,
      "eval_samples_per_second": 4.834,
      "eval_steps_per_second": 0.155,
      "step": 650
    },
    {
      "epoch": 0.17944535073409462,
      "grad_norm": 4.46875,
      "learning_rate": 3.3e-05,
      "loss": 0.357,
      "step": 660
    },
    {
      "epoch": 0.1821642196846112,
      "grad_norm": 3.3125,
      "learning_rate": 3.35e-05,
      "loss": 0.3619,
      "step": 670
    },
    {
      "epoch": 0.18488308863512778,
      "grad_norm": 4.59375,
      "learning_rate": 3.4000000000000007e-05,
      "loss": 0.3609,
      "step": 680
    },
    {
      "epoch": 0.18760195758564438,
      "grad_norm": 3.4375,
      "learning_rate": 3.45e-05,
      "loss": 0.3563,
      "step": 690
    },
    {
      "epoch": 0.19032082653616095,
      "grad_norm": 6.5,
      "learning_rate": 3.5e-05,
      "loss": 0.3225,
      "step": 700
    },
    {
      "epoch": 0.19032082653616095,
      "eval_loss": 0.33390626311302185,
      "eval_runtime": 101.5896,
      "eval_samples_per_second": 4.922,
      "eval_steps_per_second": 0.157,
      "step": 700
    },
    {
      "epoch": 0.19303969548667754,
      "grad_norm": 3.34375,
      "learning_rate": 3.55e-05,
      "loss": 0.3589,
      "step": 710
    },
    {
      "epoch": 0.19575856443719414,
      "grad_norm": 5.40625,
      "learning_rate": 3.6e-05,
      "loss": 0.3582,
      "step": 720
    },
    {
      "epoch": 0.1984774333877107,
      "grad_norm": 3.65625,
      "learning_rate": 3.65e-05,
      "loss": 0.3336,
      "step": 730
    },
    {
      "epoch": 0.2011963023382273,
      "grad_norm": 4.1875,
      "learning_rate": 3.7e-05,
      "loss": 0.2898,
      "step": 740
    },
    {
      "epoch": 0.2039151712887439,
      "grad_norm": 2.890625,
      "learning_rate": 3.7500000000000003e-05,
      "loss": 0.3009,
      "step": 750
    },
    {
      "epoch": 0.2039151712887439,
      "eval_loss": 0.30720311403274536,
      "eval_runtime": 102.4679,
      "eval_samples_per_second": 4.88,
      "eval_steps_per_second": 0.156,
      "step": 750
    },
    {
      "epoch": 0.20663404023926046,
      "grad_norm": 3.328125,
      "learning_rate": 3.8e-05,
      "loss": 0.289,
      "step": 760
    },
    {
      "epoch": 0.20935290918977706,
      "grad_norm": 2.0,
      "learning_rate": 3.85e-05,
      "loss": 0.3202,
      "step": 770
    },
    {
      "epoch": 0.21207177814029363,
      "grad_norm": 2.5,
      "learning_rate": 3.9000000000000006e-05,
      "loss": 0.2801,
      "step": 780
    },
    {
      "epoch": 0.21479064709081022,
      "grad_norm": 3.4375,
      "learning_rate": 3.9500000000000005e-05,
      "loss": 0.3109,
      "step": 790
    },
    {
      "epoch": 0.21750951604132682,
      "grad_norm": 2.046875,
      "learning_rate": 4e-05,
      "loss": 0.2771,
      "step": 800
    },
    {
      "epoch": 0.21750951604132682,
      "eval_loss": 0.2799687385559082,
      "eval_runtime": 102.1626,
      "eval_samples_per_second": 4.894,
      "eval_steps_per_second": 0.157,
      "step": 800
    },
    {
      "epoch": 0.22022838499184338,
      "grad_norm": 2.75,
      "learning_rate": 4.05e-05,
      "loss": 0.3018,
      "step": 810
    },
    {
      "epoch": 0.22294725394235998,
      "grad_norm": 2.34375,
      "learning_rate": 4.1e-05,
      "loss": 0.3067,
      "step": 820
    },
    {
      "epoch": 0.22566612289287658,
      "grad_norm": 2.9375,
      "learning_rate": 4.15e-05,
      "loss": 0.2698,
      "step": 830
    },
    {
      "epoch": 0.22838499184339314,
      "grad_norm": 3.859375,
      "learning_rate": 4.2e-05,
      "loss": 0.287,
      "step": 840
    },
    {
      "epoch": 0.23110386079390974,
      "grad_norm": 2.5625,
      "learning_rate": 4.25e-05,
      "loss": 0.2698,
      "step": 850
    },
    {
      "epoch": 0.23110386079390974,
      "eval_loss": 0.2582656145095825,
      "eval_runtime": 99.5709,
      "eval_samples_per_second": 5.022,
      "eval_steps_per_second": 0.161,
      "step": 850
    },
    {
      "epoch": 0.2338227297444263,
      "grad_norm": 1.7890625,
      "learning_rate": 4.3e-05,
      "loss": 0.2944,
      "step": 860
    },
    {
      "epoch": 0.2365415986949429,
      "grad_norm": 4.9375,
      "learning_rate": 4.35e-05,
      "loss": 0.277,
      "step": 870
    },
    {
      "epoch": 0.2392604676454595,
      "grad_norm": 3.421875,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2622,
      "step": 880
    },
    {
      "epoch": 0.24197933659597606,
      "grad_norm": 3.59375,
      "learning_rate": 4.4500000000000004e-05,
      "loss": 0.2723,
      "step": 890
    },
    {
      "epoch": 0.24469820554649266,
      "grad_norm": 2.1875,
      "learning_rate": 4.5e-05,
      "loss": 0.2674,
      "step": 900
    },
    {
      "epoch": 0.24469820554649266,
      "eval_loss": 0.26100000739097595,
      "eval_runtime": 108.8602,
      "eval_samples_per_second": 4.593,
      "eval_steps_per_second": 0.147,
      "step": 900
    },
    {
      "epoch": 0.24741707449700925,
      "grad_norm": 3.0625,
      "learning_rate": 4.55e-05,
      "loss": 0.2937,
      "step": 910
    },
    {
      "epoch": 0.2501359434475258,
      "grad_norm": 3.640625,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.2532,
      "step": 920
    },
    {
      "epoch": 0.2528548123980424,
      "grad_norm": 2.15625,
      "learning_rate": 4.6500000000000005e-05,
      "loss": 0.2252,
      "step": 930
    },
    {
      "epoch": 0.255573681348559,
      "grad_norm": 5.125,
      "learning_rate": 4.7e-05,
      "loss": 0.3014,
      "step": 940
    },
    {
      "epoch": 0.2582925502990756,
      "grad_norm": 3.140625,
      "learning_rate": 4.75e-05,
      "loss": 0.2762,
      "step": 950
    },
    {
      "epoch": 0.2582925502990756,
      "eval_loss": 0.2495156228542328,
      "eval_runtime": 102.3702,
      "eval_samples_per_second": 4.884,
      "eval_steps_per_second": 0.156,
      "step": 950
    },
    {
      "epoch": 0.26101141924959215,
      "grad_norm": 2.328125,
      "learning_rate": 4.8e-05,
      "loss": 0.2849,
      "step": 960
    },
    {
      "epoch": 0.26373028820010874,
      "grad_norm": 2.21875,
      "learning_rate": 4.85e-05,
      "loss": 0.2518,
      "step": 970
    },
    {
      "epoch": 0.26644915715062534,
      "grad_norm": 2.328125,
      "learning_rate": 4.9e-05,
      "loss": 0.2613,
      "step": 980
    },
    {
      "epoch": 0.26916802610114193,
      "grad_norm": 2.640625,
      "learning_rate": 4.9500000000000004e-05,
      "loss": 0.2867,
      "step": 990
    },
    {
      "epoch": 0.27188689505165853,
      "grad_norm": 2.3125,
      "learning_rate": 5e-05,
      "loss": 0.2569,
      "step": 1000
    },
    {
      "epoch": 0.27188689505165853,
      "eval_loss": 0.2461249977350235,
      "eval_runtime": 99.7432,
      "eval_samples_per_second": 5.013,
      "eval_steps_per_second": 0.16,
      "step": 1000
    },
    {
      "epoch": 0.27460576400217507,
      "grad_norm": 2.734375,
      "learning_rate": 4.997124784358827e-05,
      "loss": 0.2516,
      "step": 1010
    },
    {
      "epoch": 0.27732463295269166,
      "grad_norm": 3.015625,
      "learning_rate": 4.994249568717654e-05,
      "loss": 0.2258,
      "step": 1020
    },
    {
      "epoch": 0.28004350190320826,
      "grad_norm": 2.828125,
      "learning_rate": 4.991374353076481e-05,
      "loss": 0.254,
      "step": 1030
    },
    {
      "epoch": 0.28276237085372485,
      "grad_norm": 2.453125,
      "learning_rate": 4.988499137435308e-05,
      "loss": 0.2312,
      "step": 1040
    },
    {
      "epoch": 0.28548123980424145,
      "grad_norm": 2.703125,
      "learning_rate": 4.9856239217941346e-05,
      "loss": 0.2609,
      "step": 1050
    },
    {
      "epoch": 0.28548123980424145,
      "eval_loss": 0.24101562798023224,
      "eval_runtime": 101.2014,
      "eval_samples_per_second": 4.941,
      "eval_steps_per_second": 0.158,
      "step": 1050
    },
    {
      "epoch": 0.28820010875475804,
      "grad_norm": 1.875,
      "learning_rate": 4.982748706152961e-05,
      "loss": 0.263,
      "step": 1060
    },
    {
      "epoch": 0.2909189777052746,
      "grad_norm": 1.5625,
      "learning_rate": 4.979873490511789e-05,
      "loss": 0.1983,
      "step": 1070
    },
    {
      "epoch": 0.2936378466557912,
      "grad_norm": 3.09375,
      "learning_rate": 4.9769982748706154e-05,
      "loss": 0.2238,
      "step": 1080
    },
    {
      "epoch": 0.2963567156063078,
      "grad_norm": 2.21875,
      "learning_rate": 4.974123059229442e-05,
      "loss": 0.2187,
      "step": 1090
    },
    {
      "epoch": 0.29907558455682437,
      "grad_norm": 2.703125,
      "learning_rate": 4.971247843588269e-05,
      "loss": 0.2214,
      "step": 1100
    },
    {
      "epoch": 0.29907558455682437,
      "eval_loss": 0.2195468693971634,
      "eval_runtime": 98.8021,
      "eval_samples_per_second": 5.061,
      "eval_steps_per_second": 0.162,
      "step": 1100
    },
    {
      "epoch": 0.30179445350734097,
      "grad_norm": 1.4765625,
      "learning_rate": 4.968372627947096e-05,
      "loss": 0.2391,
      "step": 1110
    },
    {
      "epoch": 0.3045133224578575,
      "grad_norm": 2.640625,
      "learning_rate": 4.965497412305923e-05,
      "loss": 0.2816,
      "step": 1120
    },
    {
      "epoch": 0.3072321914083741,
      "grad_norm": 4.0,
      "learning_rate": 4.9626221966647505e-05,
      "loss": 0.2376,
      "step": 1130
    },
    {
      "epoch": 0.3099510603588907,
      "grad_norm": 2.953125,
      "learning_rate": 4.959746981023577e-05,
      "loss": 0.2369,
      "step": 1140
    },
    {
      "epoch": 0.3126699293094073,
      "grad_norm": 2.359375,
      "learning_rate": 4.956871765382404e-05,
      "loss": 0.2716,
      "step": 1150
    },
    {
      "epoch": 0.3126699293094073,
      "eval_loss": 0.2147500067949295,
      "eval_runtime": 99.2331,
      "eval_samples_per_second": 5.039,
      "eval_steps_per_second": 0.161,
      "step": 1150
    },
    {
      "epoch": 0.3153887982599239,
      "grad_norm": 1.96875,
      "learning_rate": 4.953996549741231e-05,
      "loss": 0.2294,
      "step": 1160
    },
    {
      "epoch": 0.3181076672104405,
      "grad_norm": 3.125,
      "learning_rate": 4.951121334100058e-05,
      "loss": 0.229,
      "step": 1170
    },
    {
      "epoch": 0.320826536160957,
      "grad_norm": 2.15625,
      "learning_rate": 4.948246118458885e-05,
      "loss": 0.2373,
      "step": 1180
    },
    {
      "epoch": 0.3235454051114736,
      "grad_norm": 3.859375,
      "learning_rate": 4.9453709028177115e-05,
      "loss": 0.2024,
      "step": 1190
    },
    {
      "epoch": 0.3262642740619902,
      "grad_norm": 2.984375,
      "learning_rate": 4.942495687176538e-05,
      "loss": 0.226,
      "step": 1200
    },
    {
      "epoch": 0.3262642740619902,
      "eval_loss": 0.21196874976158142,
      "eval_runtime": 97.8646,
      "eval_samples_per_second": 5.109,
      "eval_steps_per_second": 0.163,
      "step": 1200
    },
    {
      "epoch": 0.3289831430125068,
      "grad_norm": 3.59375,
      "learning_rate": 4.939620471535366e-05,
      "loss": 0.2074,
      "step": 1210
    },
    {
      "epoch": 0.3317020119630234,
      "grad_norm": 2.796875,
      "learning_rate": 4.9367452558941924e-05,
      "loss": 0.2774,
      "step": 1220
    },
    {
      "epoch": 0.33442088091353994,
      "grad_norm": 2.015625,
      "learning_rate": 4.933870040253019e-05,
      "loss": 0.2415,
      "step": 1230
    },
    {
      "epoch": 0.33713974986405654,
      "grad_norm": 2.296875,
      "learning_rate": 4.930994824611846e-05,
      "loss": 0.2433,
      "step": 1240
    },
    {
      "epoch": 0.33985861881457313,
      "grad_norm": 2.078125,
      "learning_rate": 4.928119608970673e-05,
      "loss": 0.1956,
      "step": 1250
    },
    {
      "epoch": 0.33985861881457313,
      "eval_loss": 0.2102968692779541,
      "eval_runtime": 99.6548,
      "eval_samples_per_second": 5.017,
      "eval_steps_per_second": 0.161,
      "step": 1250
    },
    {
      "epoch": 0.3425774877650897,
      "grad_norm": 2.234375,
      "learning_rate": 4.9252443933295e-05,
      "loss": 0.2314,
      "step": 1260
    },
    {
      "epoch": 0.3452963567156063,
      "grad_norm": 4.03125,
      "learning_rate": 4.922369177688327e-05,
      "loss": 0.2145,
      "step": 1270
    },
    {
      "epoch": 0.3480152256661229,
      "grad_norm": 1.6796875,
      "learning_rate": 4.9194939620471535e-05,
      "loss": 0.2047,
      "step": 1280
    },
    {
      "epoch": 0.35073409461663946,
      "grad_norm": 3.75,
      "learning_rate": 4.91661874640598e-05,
      "loss": 0.2275,
      "step": 1290
    },
    {
      "epoch": 0.35345296356715605,
      "grad_norm": 1.90625,
      "learning_rate": 4.9137435307648076e-05,
      "loss": 0.1917,
      "step": 1300
    },
    {
      "epoch": 0.35345296356715605,
      "eval_loss": 0.18660937249660492,
      "eval_runtime": 96.2132,
      "eval_samples_per_second": 5.197,
      "eval_steps_per_second": 0.166,
      "step": 1300
    },
    {
      "epoch": 0.35617183251767265,
      "grad_norm": 3.171875,
      "learning_rate": 4.910868315123634e-05,
      "loss": 0.2175,
      "step": 1310
    },
    {
      "epoch": 0.35889070146818924,
      "grad_norm": 3.03125,
      "learning_rate": 4.907993099482461e-05,
      "loss": 0.2079,
      "step": 1320
    },
    {
      "epoch": 0.36160957041870584,
      "grad_norm": 2.265625,
      "learning_rate": 4.905117883841288e-05,
      "loss": 0.2022,
      "step": 1330
    },
    {
      "epoch": 0.3643284393692224,
      "grad_norm": 1.71875,
      "learning_rate": 4.902242668200115e-05,
      "loss": 0.2294,
      "step": 1340
    },
    {
      "epoch": 0.367047308319739,
      "grad_norm": 1.6640625,
      "learning_rate": 4.899367452558942e-05,
      "loss": 0.2036,
      "step": 1350
    },
    {
      "epoch": 0.367047308319739,
      "eval_loss": 0.193296879529953,
      "eval_runtime": 98.0614,
      "eval_samples_per_second": 5.099,
      "eval_steps_per_second": 0.163,
      "step": 1350
    },
    {
      "epoch": 0.36976617727025557,
      "grad_norm": 1.671875,
      "learning_rate": 4.8964922369177693e-05,
      "loss": 0.221,
      "step": 1360
    },
    {
      "epoch": 0.37248504622077216,
      "grad_norm": 1.8359375,
      "learning_rate": 4.893617021276596e-05,
      "loss": 0.2364,
      "step": 1370
    },
    {
      "epoch": 0.37520391517128876,
      "grad_norm": 2.6875,
      "learning_rate": 4.8907418056354235e-05,
      "loss": 0.2026,
      "step": 1380
    },
    {
      "epoch": 0.37792278412180536,
      "grad_norm": 2.21875,
      "learning_rate": 4.88786658999425e-05,
      "loss": 0.1837,
      "step": 1390
    },
    {
      "epoch": 0.3806416530723219,
      "grad_norm": 1.8125,
      "learning_rate": 4.884991374353077e-05,
      "loss": 0.2146,
      "step": 1400
    },
    {
      "epoch": 0.3806416530723219,
      "eval_loss": 0.1835000067949295,
      "eval_runtime": 116.5067,
      "eval_samples_per_second": 4.292,
      "eval_steps_per_second": 0.137,
      "step": 1400
    },
    {
      "epoch": 0.3833605220228385,
      "grad_norm": 1.8046875,
      "learning_rate": 4.882116158711904e-05,
      "loss": 0.2244,
      "step": 1410
    },
    {
      "epoch": 0.3860793909733551,
      "grad_norm": 2.28125,
      "learning_rate": 4.8792409430707304e-05,
      "loss": 0.2105,
      "step": 1420
    },
    {
      "epoch": 0.3887982599238717,
      "grad_norm": 2.96875,
      "learning_rate": 4.876365727429558e-05,
      "loss": 0.244,
      "step": 1430
    },
    {
      "epoch": 0.3915171288743883,
      "grad_norm": 1.3984375,
      "learning_rate": 4.8734905117883846e-05,
      "loss": 0.1802,
      "step": 1440
    },
    {
      "epoch": 0.3942359978249048,
      "grad_norm": 1.71875,
      "learning_rate": 4.870615296147211e-05,
      "loss": 0.179,
      "step": 1450
    },
    {
      "epoch": 0.3942359978249048,
      "eval_loss": 0.18214061856269836,
      "eval_runtime": 104.8925,
      "eval_samples_per_second": 4.767,
      "eval_steps_per_second": 0.153,
      "step": 1450
    },
    {
      "epoch": 0.3969548667754214,
      "grad_norm": 2.421875,
      "learning_rate": 4.867740080506038e-05,
      "loss": 0.2356,
      "step": 1460
    },
    {
      "epoch": 0.399673735725938,
      "grad_norm": 1.640625,
      "learning_rate": 4.8648648648648654e-05,
      "loss": 0.2344,
      "step": 1470
    },
    {
      "epoch": 0.4023926046764546,
      "grad_norm": 1.90625,
      "learning_rate": 4.861989649223692e-05,
      "loss": 0.2065,
      "step": 1480
    },
    {
      "epoch": 0.4051114736269712,
      "grad_norm": 2.984375,
      "learning_rate": 4.859114433582519e-05,
      "loss": 0.2208,
      "step": 1490
    },
    {
      "epoch": 0.4078303425774878,
      "grad_norm": 1.4375,
      "learning_rate": 4.8562392179413456e-05,
      "loss": 0.1798,
      "step": 1500
    },
    {
      "epoch": 0.4078303425774878,
      "eval_loss": 0.18092188239097595,
      "eval_runtime": 102.421,
      "eval_samples_per_second": 4.882,
      "eval_steps_per_second": 0.156,
      "step": 1500
    },
    {
      "epoch": 0.41054921152800433,
      "grad_norm": 3.1875,
      "learning_rate": 4.8533640023001724e-05,
      "loss": 0.2177,
      "step": 1510
    },
    {
      "epoch": 0.4132680804785209,
      "grad_norm": 2.234375,
      "learning_rate": 4.850488786659e-05,
      "loss": 0.1699,
      "step": 1520
    },
    {
      "epoch": 0.4159869494290375,
      "grad_norm": 1.6171875,
      "learning_rate": 4.8476135710178265e-05,
      "loss": 0.2431,
      "step": 1530
    },
    {
      "epoch": 0.4187058183795541,
      "grad_norm": 2.5,
      "learning_rate": 4.844738355376653e-05,
      "loss": 0.2206,
      "step": 1540
    },
    {
      "epoch": 0.4214246873300707,
      "grad_norm": 2.15625,
      "learning_rate": 4.84186313973548e-05,
      "loss": 0.2051,
      "step": 1550
    },
    {
      "epoch": 0.4214246873300707,
      "eval_loss": 0.18370312452316284,
      "eval_runtime": 102.2352,
      "eval_samples_per_second": 4.891,
      "eval_steps_per_second": 0.157,
      "step": 1550
    },
    {
      "epoch": 0.42414355628058725,
      "grad_norm": 9.4375,
      "learning_rate": 4.8389879240943074e-05,
      "loss": 0.2279,
      "step": 1560
    },
    {
      "epoch": 0.42686242523110385,
      "grad_norm": 4.84375,
      "learning_rate": 4.836112708453134e-05,
      "loss": 0.2075,
      "step": 1570
    },
    {
      "epoch": 0.42958129418162044,
      "grad_norm": 1.625,
      "learning_rate": 4.833237492811961e-05,
      "loss": 0.1999,
      "step": 1580
    },
    {
      "epoch": 0.43230016313213704,
      "grad_norm": 3.421875,
      "learning_rate": 4.830362277170788e-05,
      "loss": 0.2313,
      "step": 1590
    },
    {
      "epoch": 0.43501903208265363,
      "grad_norm": 1.8359375,
      "learning_rate": 4.827487061529615e-05,
      "loss": 0.1905,
      "step": 1600
    },
    {
      "epoch": 0.43501903208265363,
      "eval_loss": 0.1731249988079071,
      "eval_runtime": 101.4653,
      "eval_samples_per_second": 4.928,
      "eval_steps_per_second": 0.158,
      "step": 1600
    },
    {
      "epoch": 0.4377379010331702,
      "grad_norm": 2.78125,
      "learning_rate": 4.8246118458884424e-05,
      "loss": 0.2297,
      "step": 1610
    },
    {
      "epoch": 0.44045676998368677,
      "grad_norm": 2.03125,
      "learning_rate": 4.821736630247269e-05,
      "loss": 0.2167,
      "step": 1620
    },
    {
      "epoch": 0.44317563893420336,
      "grad_norm": 2.0,
      "learning_rate": 4.818861414606096e-05,
      "loss": 0.2212,
      "step": 1630
    },
    {
      "epoch": 0.44589450788471996,
      "grad_norm": 1.9296875,
      "learning_rate": 4.8159861989649226e-05,
      "loss": 0.1963,
      "step": 1640
    },
    {
      "epoch": 0.44861337683523655,
      "grad_norm": 2.28125,
      "learning_rate": 4.81311098332375e-05,
      "loss": 0.1932,
      "step": 1650
    },
    {
      "epoch": 0.44861337683523655,
      "eval_loss": 0.18182812631130219,
      "eval_runtime": 101.6213,
      "eval_samples_per_second": 4.92,
      "eval_steps_per_second": 0.157,
      "step": 1650
    },
    {
      "epoch": 0.45133224578575315,
      "grad_norm": 2.640625,
      "learning_rate": 4.810235767682577e-05,
      "loss": 0.2038,
      "step": 1660
    },
    {
      "epoch": 0.4540511147362697,
      "grad_norm": 1.1484375,
      "learning_rate": 4.8073605520414034e-05,
      "loss": 0.1932,
      "step": 1670
    },
    {
      "epoch": 0.4567699836867863,
      "grad_norm": 1.984375,
      "learning_rate": 4.80448533640023e-05,
      "loss": 0.1926,
      "step": 1680
    },
    {
      "epoch": 0.4594888526373029,
      "grad_norm": 1.1328125,
      "learning_rate": 4.801610120759057e-05,
      "loss": 0.2065,
      "step": 1690
    },
    {
      "epoch": 0.4622077215878195,
      "grad_norm": 1.75,
      "learning_rate": 4.798734905117884e-05,
      "loss": 0.1948,
      "step": 1700
    },
    {
      "epoch": 0.4622077215878195,
      "eval_loss": 0.16239061951637268,
      "eval_runtime": 100.5121,
      "eval_samples_per_second": 4.975,
      "eval_steps_per_second": 0.159,
      "step": 1700
    },
    {
      "epoch": 0.46492659053833607,
      "grad_norm": 1.7734375,
      "learning_rate": 4.795859689476711e-05,
      "loss": 0.2052,
      "step": 1710
    },
    {
      "epoch": 0.4676454594888526,
      "grad_norm": 2.046875,
      "learning_rate": 4.792984473835538e-05,
      "loss": 0.2364,
      "step": 1720
    },
    {
      "epoch": 0.4703643284393692,
      "grad_norm": 2.078125,
      "learning_rate": 4.7901092581943645e-05,
      "loss": 0.1674,
      "step": 1730
    },
    {
      "epoch": 0.4730831973898858,
      "grad_norm": 2.78125,
      "learning_rate": 4.787234042553192e-05,
      "loss": 0.1733,
      "step": 1740
    },
    {
      "epoch": 0.4758020663404024,
      "grad_norm": 2.765625,
      "learning_rate": 4.7843588269120187e-05,
      "loss": 0.2512,
      "step": 1750
    },
    {
      "epoch": 0.4758020663404024,
      "eval_loss": 0.1744687557220459,
      "eval_runtime": 98.5091,
      "eval_samples_per_second": 5.076,
      "eval_steps_per_second": 0.162,
      "step": 1750
    },
    {
      "epoch": 0.478520935290919,
      "grad_norm": 1.9140625,
      "learning_rate": 4.7814836112708454e-05,
      "loss": 0.2165,
      "step": 1760
    },
    {
      "epoch": 0.4812398042414356,
      "grad_norm": 1.765625,
      "learning_rate": 4.778608395629672e-05,
      "loss": 0.1927,
      "step": 1770
    },
    {
      "epoch": 0.4839586731919521,
      "grad_norm": 1.7734375,
      "learning_rate": 4.775733179988499e-05,
      "loss": 0.195,
      "step": 1780
    },
    {
      "epoch": 0.4866775421424687,
      "grad_norm": 1.8359375,
      "learning_rate": 4.772857964347326e-05,
      "loss": 0.1674,
      "step": 1790
    },
    {
      "epoch": 0.4893964110929853,
      "grad_norm": 2.234375,
      "learning_rate": 4.769982748706153e-05,
      "loss": 0.1971,
      "step": 1800
    },
    {
      "epoch": 0.4893964110929853,
      "eval_loss": 0.16629686951637268,
      "eval_runtime": 102.338,
      "eval_samples_per_second": 4.886,
      "eval_steps_per_second": 0.156,
      "step": 1800
    },
    {
      "epoch": 0.4921152800435019,
      "grad_norm": 1.734375,
      "learning_rate": 4.76710753306498e-05,
      "loss": 0.1645,
      "step": 1810
    },
    {
      "epoch": 0.4948341489940185,
      "grad_norm": 7.375,
      "learning_rate": 4.764232317423807e-05,
      "loss": 0.2437,
      "step": 1820
    },
    {
      "epoch": 0.49755301794453505,
      "grad_norm": 2.859375,
      "learning_rate": 4.761357101782634e-05,
      "loss": 0.152,
      "step": 1830
    },
    {
      "epoch": 0.5002718868950516,
      "grad_norm": 3.640625,
      "learning_rate": 4.758481886141461e-05,
      "loss": 0.1832,
      "step": 1840
    },
    {
      "epoch": 0.5029907558455683,
      "grad_norm": 2.28125,
      "learning_rate": 4.755606670500288e-05,
      "loss": 0.1836,
      "step": 1850
    },
    {
      "epoch": 0.5029907558455683,
      "eval_loss": 0.16115625202655792,
      "eval_runtime": 99.419,
      "eval_samples_per_second": 5.029,
      "eval_steps_per_second": 0.161,
      "step": 1850
    },
    {
      "epoch": 0.5057096247960848,
      "grad_norm": 1.8984375,
      "learning_rate": 4.752731454859115e-05,
      "loss": 0.1636,
      "step": 1860
    },
    {
      "epoch": 0.5084284937466014,
      "grad_norm": 1.734375,
      "learning_rate": 4.7498562392179415e-05,
      "loss": 0.211,
      "step": 1870
    },
    {
      "epoch": 0.511147362697118,
      "grad_norm": 1.125,
      "learning_rate": 4.746981023576769e-05,
      "loss": 0.1754,
      "step": 1880
    },
    {
      "epoch": 0.5138662316476346,
      "grad_norm": 1.1328125,
      "learning_rate": 4.7441058079355956e-05,
      "loss": 0.1891,
      "step": 1890
    },
    {
      "epoch": 0.5165851005981512,
      "grad_norm": 2.421875,
      "learning_rate": 4.741230592294422e-05,
      "loss": 0.1958,
      "step": 1900
    },
    {
      "epoch": 0.5165851005981512,
      "eval_loss": 0.16267187893390656,
      "eval_runtime": 118.6656,
      "eval_samples_per_second": 4.214,
      "eval_steps_per_second": 0.135,
      "step": 1900
    },
    {
      "epoch": 0.5193039695486678,
      "grad_norm": 1.15625,
      "learning_rate": 4.738355376653249e-05,
      "loss": 0.1745,
      "step": 1910
    },
    {
      "epoch": 0.5220228384991843,
      "grad_norm": 1.625,
      "learning_rate": 4.7354801610120765e-05,
      "loss": 0.1658,
      "step": 1920
    },
    {
      "epoch": 0.524741707449701,
      "grad_norm": 2.140625,
      "learning_rate": 4.732604945370903e-05,
      "loss": 0.1675,
      "step": 1930
    },
    {
      "epoch": 0.5274605764002175,
      "grad_norm": 1.4453125,
      "learning_rate": 4.72972972972973e-05,
      "loss": 0.1814,
      "step": 1940
    },
    {
      "epoch": 0.5301794453507341,
      "grad_norm": 2.71875,
      "learning_rate": 4.726854514088557e-05,
      "loss": 0.1899,
      "step": 1950
    },
    {
      "epoch": 0.5301794453507341,
      "eval_loss": 0.16292187571525574,
      "eval_runtime": 107.7806,
      "eval_samples_per_second": 4.639,
      "eval_steps_per_second": 0.148,
      "step": 1950
    },
    {
      "epoch": 0.5328983143012507,
      "grad_norm": 1.8984375,
      "learning_rate": 4.7239792984473834e-05,
      "loss": 0.1981,
      "step": 1960
    },
    {
      "epoch": 0.5356171832517672,
      "grad_norm": 2.015625,
      "learning_rate": 4.721104082806211e-05,
      "loss": 0.1706,
      "step": 1970
    },
    {
      "epoch": 0.5383360522022839,
      "grad_norm": 3.265625,
      "learning_rate": 4.7182288671650375e-05,
      "loss": 0.1958,
      "step": 1980
    },
    {
      "epoch": 0.5410549211528004,
      "grad_norm": 2.9375,
      "learning_rate": 4.715353651523864e-05,
      "loss": 0.2111,
      "step": 1990
    },
    {
      "epoch": 0.5437737901033171,
      "grad_norm": 3.484375,
      "learning_rate": 4.712478435882691e-05,
      "loss": 0.1983,
      "step": 2000
    },
    {
      "epoch": 0.5437737901033171,
      "eval_loss": 0.16431249678134918,
      "eval_runtime": 104.4508,
      "eval_samples_per_second": 4.787,
      "eval_steps_per_second": 0.153,
      "step": 2000
    },
    {
      "epoch": 0.5464926590538336,
      "grad_norm": 1.71875,
      "learning_rate": 4.7096032202415184e-05,
      "loss": 0.1587,
      "step": 2010
    },
    {
      "epoch": 0.5492115280043501,
      "grad_norm": 1.7734375,
      "learning_rate": 4.706728004600345e-05,
      "loss": 0.1746,
      "step": 2020
    },
    {
      "epoch": 0.5519303969548668,
      "grad_norm": 2.15625,
      "learning_rate": 4.703852788959172e-05,
      "loss": 0.1666,
      "step": 2030
    },
    {
      "epoch": 0.5546492659053833,
      "grad_norm": 1.859375,
      "learning_rate": 4.7009775733179986e-05,
      "loss": 0.1779,
      "step": 2040
    },
    {
      "epoch": 0.5573681348559,
      "grad_norm": 2.53125,
      "learning_rate": 4.698102357676826e-05,
      "loss": 0.2026,
      "step": 2050
    },
    {
      "epoch": 0.5573681348559,
      "eval_loss": 0.1557968705892563,
      "eval_runtime": 104.3041,
      "eval_samples_per_second": 4.794,
      "eval_steps_per_second": 0.153,
      "step": 2050
    },
    {
      "epoch": 0.5600870038064165,
      "grad_norm": 2.578125,
      "learning_rate": 4.695227142035653e-05,
      "loss": 0.1878,
      "step": 2060
    },
    {
      "epoch": 0.5628058727569332,
      "grad_norm": 1.0234375,
      "learning_rate": 4.69235192639448e-05,
      "loss": 0.1482,
      "step": 2070
    },
    {
      "epoch": 0.5655247417074497,
      "grad_norm": 1.421875,
      "learning_rate": 4.689476710753307e-05,
      "loss": 0.1903,
      "step": 2080
    },
    {
      "epoch": 0.5682436106579662,
      "grad_norm": 2.75,
      "learning_rate": 4.6866014951121336e-05,
      "loss": 0.1969,
      "step": 2090
    },
    {
      "epoch": 0.5709624796084829,
      "grad_norm": 1.6875,
      "learning_rate": 4.683726279470961e-05,
      "loss": 0.1938,
      "step": 2100
    },
    {
      "epoch": 0.5709624796084829,
      "eval_loss": 0.15632812678813934,
      "eval_runtime": 100.4891,
      "eval_samples_per_second": 4.976,
      "eval_steps_per_second": 0.159,
      "step": 2100
    },
    {
      "epoch": 0.5736813485589994,
      "grad_norm": 2.203125,
      "learning_rate": 4.680851063829788e-05,
      "loss": 0.177,
      "step": 2110
    },
    {
      "epoch": 0.5764002175095161,
      "grad_norm": 1.484375,
      "learning_rate": 4.6779758481886145e-05,
      "loss": 0.1903,
      "step": 2120
    },
    {
      "epoch": 0.5791190864600326,
      "grad_norm": 0.9140625,
      "learning_rate": 4.675100632547441e-05,
      "loss": 0.2083,
      "step": 2130
    },
    {
      "epoch": 0.5818379554105492,
      "grad_norm": 1.4921875,
      "learning_rate": 4.6722254169062686e-05,
      "loss": 0.1624,
      "step": 2140
    },
    {
      "epoch": 0.5845568243610658,
      "grad_norm": 2.0625,
      "learning_rate": 4.6693502012650954e-05,
      "loss": 0.1713,
      "step": 2150
    },
    {
      "epoch": 0.5845568243610658,
      "eval_loss": 0.16362500190734863,
      "eval_runtime": 98.6909,
      "eval_samples_per_second": 5.066,
      "eval_steps_per_second": 0.162,
      "step": 2150
    },
    {
      "epoch": 0.5872756933115824,
      "grad_norm": 1.265625,
      "learning_rate": 4.666474985623922e-05,
      "loss": 0.1722,
      "step": 2160
    },
    {
      "epoch": 0.589994562262099,
      "grad_norm": 2.109375,
      "learning_rate": 4.663599769982749e-05,
      "loss": 0.1798,
      "step": 2170
    },
    {
      "epoch": 0.5927134312126155,
      "grad_norm": 1.6875,
      "learning_rate": 4.6607245543415756e-05,
      "loss": 0.1344,
      "step": 2180
    },
    {
      "epoch": 0.5954323001631321,
      "grad_norm": 1.59375,
      "learning_rate": 4.657849338700403e-05,
      "loss": 0.1726,
      "step": 2190
    },
    {
      "epoch": 0.5981511691136487,
      "grad_norm": 1.1796875,
      "learning_rate": 4.65497412305923e-05,
      "loss": 0.1564,
      "step": 2200
    },
    {
      "epoch": 0.5981511691136487,
      "eval_loss": 0.15471875667572021,
      "eval_runtime": 98.3773,
      "eval_samples_per_second": 5.082,
      "eval_steps_per_second": 0.163,
      "step": 2200
    },
    {
      "epoch": 0.6008700380641653,
      "grad_norm": 1.5703125,
      "learning_rate": 4.6520989074180564e-05,
      "loss": 0.1767,
      "step": 2210
    },
    {
      "epoch": 0.6035889070146819,
      "grad_norm": 1.65625,
      "learning_rate": 4.649223691776883e-05,
      "loss": 0.1828,
      "step": 2220
    },
    {
      "epoch": 0.6063077759651985,
      "grad_norm": 2.375,
      "learning_rate": 4.6463484761357106e-05,
      "loss": 0.1684,
      "step": 2230
    },
    {
      "epoch": 0.609026644915715,
      "grad_norm": 1.890625,
      "learning_rate": 4.643473260494537e-05,
      "loss": 0.1838,
      "step": 2240
    },
    {
      "epoch": 0.6117455138662317,
      "grad_norm": 2.015625,
      "learning_rate": 4.640598044853364e-05,
      "loss": 0.1637,
      "step": 2250
    },
    {
      "epoch": 0.6117455138662317,
      "eval_loss": 0.16246874630451202,
      "eval_runtime": 99.6916,
      "eval_samples_per_second": 5.015,
      "eval_steps_per_second": 0.16,
      "step": 2250
    },
    {
      "epoch": 0.6144643828167482,
      "grad_norm": 2.5625,
      "learning_rate": 4.637722829212191e-05,
      "loss": 0.1988,
      "step": 2260
    },
    {
      "epoch": 0.6171832517672649,
      "grad_norm": 1.828125,
      "learning_rate": 4.6348476135710175e-05,
      "loss": 0.1993,
      "step": 2270
    },
    {
      "epoch": 0.6199021207177814,
      "grad_norm": 2.515625,
      "learning_rate": 4.631972397929845e-05,
      "loss": 0.21,
      "step": 2280
    },
    {
      "epoch": 0.622620989668298,
      "grad_norm": 2.515625,
      "learning_rate": 4.6290971822886716e-05,
      "loss": 0.1944,
      "step": 2290
    },
    {
      "epoch": 0.6253398586188146,
      "grad_norm": 3.734375,
      "learning_rate": 4.626221966647499e-05,
      "loss": 0.1894,
      "step": 2300
    },
    {
      "epoch": 0.6253398586188146,
      "eval_loss": 0.15801562368869781,
      "eval_runtime": 97.8482,
      "eval_samples_per_second": 5.11,
      "eval_steps_per_second": 0.164,
      "step": 2300
    },
    {
      "epoch": 0.6280587275693311,
      "grad_norm": 2.390625,
      "learning_rate": 4.623346751006326e-05,
      "loss": 0.1735,
      "step": 2310
    },
    {
      "epoch": 0.6307775965198478,
      "grad_norm": 2.84375,
      "learning_rate": 4.620471535365153e-05,
      "loss": 0.1892,
      "step": 2320
    },
    {
      "epoch": 0.6334964654703643,
      "grad_norm": 1.875,
      "learning_rate": 4.61759631972398e-05,
      "loss": 0.1955,
      "step": 2330
    },
    {
      "epoch": 0.636215334420881,
      "grad_norm": 2.46875,
      "learning_rate": 4.6147211040828067e-05,
      "loss": 0.1777,
      "step": 2340
    },
    {
      "epoch": 0.6389342033713975,
      "grad_norm": 1.7734375,
      "learning_rate": 4.6118458884416334e-05,
      "loss": 0.1637,
      "step": 2350
    },
    {
      "epoch": 0.6389342033713975,
      "eval_loss": 0.15985937416553497,
      "eval_runtime": 109.048,
      "eval_samples_per_second": 4.585,
      "eval_steps_per_second": 0.147,
      "step": 2350
    },
    {
      "epoch": 0.641653072321914,
      "grad_norm": 0.82421875,
      "learning_rate": 4.60897067280046e-05,
      "loss": 0.1771,
      "step": 2360
    },
    {
      "epoch": 0.6443719412724307,
      "grad_norm": 1.484375,
      "learning_rate": 4.6060954571592875e-05,
      "loss": 0.1768,
      "step": 2370
    },
    {
      "epoch": 0.6470908102229472,
      "grad_norm": 1.8671875,
      "learning_rate": 4.603220241518114e-05,
      "loss": 0.1798,
      "step": 2380
    },
    {
      "epoch": 0.6498096791734639,
      "grad_norm": 1.984375,
      "learning_rate": 4.600345025876941e-05,
      "loss": 0.1535,
      "step": 2390
    },
    {
      "epoch": 0.6525285481239804,
      "grad_norm": 4.5,
      "learning_rate": 4.597469810235768e-05,
      "loss": 0.2031,
      "step": 2400
    },
    {
      "epoch": 0.6525285481239804,
      "eval_loss": 0.15268750488758087,
      "eval_runtime": 168.1057,
      "eval_samples_per_second": 2.974,
      "eval_steps_per_second": 0.095,
      "step": 2400
    },
    {
      "epoch": 0.655247417074497,
      "grad_norm": 2.015625,
      "learning_rate": 4.594594594594595e-05,
      "loss": 0.1985,
      "step": 2410
    },
    {
      "epoch": 0.6579662860250136,
      "grad_norm": 1.2890625,
      "learning_rate": 4.591719378953422e-05,
      "loss": 0.1459,
      "step": 2420
    },
    {
      "epoch": 0.6606851549755302,
      "grad_norm": 0.859375,
      "learning_rate": 4.5888441633122486e-05,
      "loss": 0.1607,
      "step": 2430
    },
    {
      "epoch": 0.6634040239260468,
      "grad_norm": 3.234375,
      "learning_rate": 4.585968947671075e-05,
      "loss": 0.2168,
      "step": 2440
    },
    {
      "epoch": 0.6661228928765633,
      "grad_norm": 2.40625,
      "learning_rate": 4.583093732029902e-05,
      "loss": 0.1696,
      "step": 2450
    },
    {
      "epoch": 0.6661228928765633,
      "eval_loss": 0.14201562106609344,
      "eval_runtime": 103.7059,
      "eval_samples_per_second": 4.821,
      "eval_steps_per_second": 0.154,
      "step": 2450
    },
    {
      "epoch": 0.6688417618270799,
      "grad_norm": 1.453125,
      "learning_rate": 4.5802185163887295e-05,
      "loss": 0.1676,
      "step": 2460
    },
    {
      "epoch": 0.6715606307775965,
      "grad_norm": 1.015625,
      "learning_rate": 4.577343300747556e-05,
      "loss": 0.1718,
      "step": 2470
    },
    {
      "epoch": 0.6742794997281131,
      "grad_norm": 1.328125,
      "learning_rate": 4.574468085106383e-05,
      "loss": 0.1561,
      "step": 2480
    },
    {
      "epoch": 0.6769983686786297,
      "grad_norm": 2.375,
      "learning_rate": 4.5715928694652097e-05,
      "loss": 0.1687,
      "step": 2490
    },
    {
      "epoch": 0.6797172376291463,
      "grad_norm": 2.03125,
      "learning_rate": 4.568717653824037e-05,
      "loss": 0.1862,
      "step": 2500
    },
    {
      "epoch": 0.6797172376291463,
      "eval_loss": 0.15104687213897705,
      "eval_runtime": 122.5851,
      "eval_samples_per_second": 4.079,
      "eval_steps_per_second": 0.131,
      "step": 2500
    },
    {
      "epoch": 0.6824361065796629,
      "grad_norm": 1.96875,
      "learning_rate": 4.565842438182864e-05,
      "loss": 0.1368,
      "step": 2510
    },
    {
      "epoch": 0.6851549755301795,
      "grad_norm": 1.25,
      "learning_rate": 4.5629672225416905e-05,
      "loss": 0.1738,
      "step": 2520
    },
    {
      "epoch": 0.687873844480696,
      "grad_norm": 2.328125,
      "learning_rate": 4.560092006900518e-05,
      "loss": 0.1688,
      "step": 2530
    },
    {
      "epoch": 0.6905927134312126,
      "grad_norm": 2.40625,
      "learning_rate": 4.557216791259345e-05,
      "loss": 0.1595,
      "step": 2540
    },
    {
      "epoch": 0.6933115823817292,
      "grad_norm": 3.390625,
      "learning_rate": 4.554341575618172e-05,
      "loss": 0.1695,
      "step": 2550
    },
    {
      "epoch": 0.6933115823817292,
      "eval_loss": 0.14746874570846558,
      "eval_runtime": 128.4375,
      "eval_samples_per_second": 3.893,
      "eval_steps_per_second": 0.125,
      "step": 2550
    },
    {
      "epoch": 0.6960304513322458,
      "grad_norm": 2.203125,
      "learning_rate": 4.551466359976999e-05,
      "loss": 0.1905,
      "step": 2560
    },
    {
      "epoch": 0.6987493202827624,
      "grad_norm": 1.5390625,
      "learning_rate": 4.5485911443358255e-05,
      "loss": 0.1637,
      "step": 2570
    },
    {
      "epoch": 0.7014681892332789,
      "grad_norm": 2.796875,
      "learning_rate": 4.545715928694652e-05,
      "loss": 0.1365,
      "step": 2580
    },
    {
      "epoch": 0.7041870581837956,
      "grad_norm": 1.578125,
      "learning_rate": 4.54284071305348e-05,
      "loss": 0.1562,
      "step": 2590
    },
    {
      "epoch": 0.7069059271343121,
      "grad_norm": 1.7421875,
      "learning_rate": 4.5399654974123064e-05,
      "loss": 0.1853,
      "step": 2600
    },
    {
      "epoch": 0.7069059271343121,
      "eval_loss": 0.15376561880111694,
      "eval_runtime": 103.9837,
      "eval_samples_per_second": 4.808,
      "eval_steps_per_second": 0.154,
      "step": 2600
    },
    {
      "epoch": 0.7096247960848288,
      "grad_norm": 1.8828125,
      "learning_rate": 4.537090281771133e-05,
      "loss": 0.1377,
      "step": 2610
    },
    {
      "epoch": 0.7123436650353453,
      "grad_norm": 1.46875,
      "learning_rate": 4.53421506612996e-05,
      "loss": 0.161,
      "step": 2620
    },
    {
      "epoch": 0.7150625339858618,
      "grad_norm": 2.890625,
      "learning_rate": 4.5313398504887866e-05,
      "loss": 0.1537,
      "step": 2630
    },
    {
      "epoch": 0.7177814029363785,
      "grad_norm": 1.9296875,
      "learning_rate": 4.528464634847614e-05,
      "loss": 0.1568,
      "step": 2640
    },
    {
      "epoch": 0.720500271886895,
      "grad_norm": 1.8671875,
      "learning_rate": 4.525589419206441e-05,
      "loss": 0.1788,
      "step": 2650
    },
    {
      "epoch": 0.720500271886895,
      "eval_loss": 0.15562500059604645,
      "eval_runtime": 102.6335,
      "eval_samples_per_second": 4.872,
      "eval_steps_per_second": 0.156,
      "step": 2650
    },
    {
      "epoch": 0.7232191408374117,
      "grad_norm": 2.8125,
      "learning_rate": 4.5227142035652675e-05,
      "loss": 0.1704,
      "step": 2660
    },
    {
      "epoch": 0.7259380097879282,
      "grad_norm": 1.1640625,
      "learning_rate": 4.519838987924094e-05,
      "loss": 0.1412,
      "step": 2670
    },
    {
      "epoch": 0.7286568787384448,
      "grad_norm": 1.34375,
      "learning_rate": 4.5169637722829216e-05,
      "loss": 0.1473,
      "step": 2680
    },
    {
      "epoch": 0.7313757476889614,
      "grad_norm": 2.25,
      "learning_rate": 4.5140885566417484e-05,
      "loss": 0.1419,
      "step": 2690
    },
    {
      "epoch": 0.734094616639478,
      "grad_norm": 1.015625,
      "learning_rate": 4.511213341000575e-05,
      "loss": 0.1598,
      "step": 2700
    },
    {
      "epoch": 0.734094616639478,
      "eval_loss": 0.15793749690055847,
      "eval_runtime": 100.5703,
      "eval_samples_per_second": 4.972,
      "eval_steps_per_second": 0.159,
      "step": 2700
    },
    {
      "epoch": 0.7368134855899946,
      "grad_norm": 1.1484375,
      "learning_rate": 4.508338125359402e-05,
      "loss": 0.1896,
      "step": 2710
    },
    {
      "epoch": 0.7395323545405111,
      "grad_norm": 4.90625,
      "learning_rate": 4.5054629097182285e-05,
      "loss": 0.173,
      "step": 2720
    },
    {
      "epoch": 0.7422512234910277,
      "grad_norm": 1.6328125,
      "learning_rate": 4.502587694077056e-05,
      "loss": 0.1781,
      "step": 2730
    },
    {
      "epoch": 0.7449700924415443,
      "grad_norm": 2.109375,
      "learning_rate": 4.499712478435883e-05,
      "loss": 0.1524,
      "step": 2740
    },
    {
      "epoch": 0.7476889613920609,
      "grad_norm": 1.578125,
      "learning_rate": 4.4968372627947094e-05,
      "loss": 0.1487,
      "step": 2750
    },
    {
      "epoch": 0.7476889613920609,
      "eval_loss": 0.1424531191587448,
      "eval_runtime": 100.4008,
      "eval_samples_per_second": 4.98,
      "eval_steps_per_second": 0.159,
      "step": 2750
    },
    {
      "epoch": 0.7504078303425775,
      "grad_norm": 2.640625,
      "learning_rate": 4.493962047153537e-05,
      "loss": 0.1978,
      "step": 2760
    },
    {
      "epoch": 0.7531266992930941,
      "grad_norm": 1.984375,
      "learning_rate": 4.4910868315123636e-05,
      "loss": 0.2137,
      "step": 2770
    },
    {
      "epoch": 0.7558455682436107,
      "grad_norm": 1.203125,
      "learning_rate": 4.488211615871191e-05,
      "loss": 0.1886,
      "step": 2780
    },
    {
      "epoch": 0.7585644371941273,
      "grad_norm": 2.078125,
      "learning_rate": 4.485336400230018e-05,
      "loss": 0.1458,
      "step": 2790
    },
    {
      "epoch": 0.7612833061446438,
      "grad_norm": 2.65625,
      "learning_rate": 4.4824611845888444e-05,
      "loss": 0.1912,
      "step": 2800
    },
    {
      "epoch": 0.7612833061446438,
      "eval_loss": 0.13567186892032623,
      "eval_runtime": 101.2689,
      "eval_samples_per_second": 4.937,
      "eval_steps_per_second": 0.158,
      "step": 2800
    },
    {
      "epoch": 0.7640021750951604,
      "grad_norm": 2.09375,
      "learning_rate": 4.479585968947671e-05,
      "loss": 0.153,
      "step": 2810
    },
    {
      "epoch": 0.766721044045677,
      "grad_norm": 1.2109375,
      "learning_rate": 4.4767107533064986e-05,
      "loss": 0.1305,
      "step": 2820
    },
    {
      "epoch": 0.7694399129961936,
      "grad_norm": 1.1015625,
      "learning_rate": 4.473835537665325e-05,
      "loss": 0.156,
      "step": 2830
    },
    {
      "epoch": 0.7721587819467102,
      "grad_norm": 2.125,
      "learning_rate": 4.470960322024152e-05,
      "loss": 0.1612,
      "step": 2840
    },
    {
      "epoch": 0.7748776508972267,
      "grad_norm": 1.4375,
      "learning_rate": 4.468085106382979e-05,
      "loss": 0.163,
      "step": 2850
    },
    {
      "epoch": 0.7748776508972267,
      "eval_loss": 0.15159374475479126,
      "eval_runtime": 109.8758,
      "eval_samples_per_second": 4.551,
      "eval_steps_per_second": 0.146,
      "step": 2850
    },
    {
      "epoch": 0.7775965198477434,
      "grad_norm": 3.171875,
      "learning_rate": 4.465209890741806e-05,
      "loss": 0.1512,
      "step": 2860
    },
    {
      "epoch": 0.7803153887982599,
      "grad_norm": 1.5234375,
      "learning_rate": 4.462334675100633e-05,
      "loss": 0.1792,
      "step": 2870
    },
    {
      "epoch": 0.7830342577487766,
      "grad_norm": 1.9296875,
      "learning_rate": 4.4594594594594596e-05,
      "loss": 0.138,
      "step": 2880
    },
    {
      "epoch": 0.7857531266992931,
      "grad_norm": 3.125,
      "learning_rate": 4.4565842438182864e-05,
      "loss": 0.171,
      "step": 2890
    },
    {
      "epoch": 0.7884719956498096,
      "grad_norm": 2.390625,
      "learning_rate": 4.453709028177114e-05,
      "loss": 0.1595,
      "step": 2900
    },
    {
      "epoch": 0.7884719956498096,
      "eval_loss": 0.1458437442779541,
      "eval_runtime": 104.1621,
      "eval_samples_per_second": 4.8,
      "eval_steps_per_second": 0.154,
      "step": 2900
    },
    {
      "epoch": 0.7911908646003263,
      "grad_norm": 1.859375,
      "learning_rate": 4.4508338125359405e-05,
      "loss": 0.206,
      "step": 2910
    },
    {
      "epoch": 0.7939097335508428,
      "grad_norm": 1.3671875,
      "learning_rate": 4.447958596894767e-05,
      "loss": 0.1424,
      "step": 2920
    },
    {
      "epoch": 0.7966286025013595,
      "grad_norm": 2.203125,
      "learning_rate": 4.445083381253594e-05,
      "loss": 0.1663,
      "step": 2930
    },
    {
      "epoch": 0.799347471451876,
      "grad_norm": 1.578125,
      "learning_rate": 4.442208165612421e-05,
      "loss": 0.1766,
      "step": 2940
    },
    {
      "epoch": 0.8020663404023926,
      "grad_norm": 1.203125,
      "learning_rate": 4.439332949971248e-05,
      "loss": 0.1588,
      "step": 2950
    },
    {
      "epoch": 0.8020663404023926,
      "eval_loss": 0.1557031273841858,
      "eval_runtime": 101.7696,
      "eval_samples_per_second": 4.913,
      "eval_steps_per_second": 0.157,
      "step": 2950
    },
    {
      "epoch": 0.8047852093529092,
      "grad_norm": 1.3828125,
      "learning_rate": 4.436457734330075e-05,
      "loss": 0.1494,
      "step": 2960
    },
    {
      "epoch": 0.8075040783034257,
      "grad_norm": 2.09375,
      "learning_rate": 4.4335825186889016e-05,
      "loss": 0.125,
      "step": 2970
    },
    {
      "epoch": 0.8102229472539424,
      "grad_norm": 1.71875,
      "learning_rate": 4.430707303047728e-05,
      "loss": 0.1642,
      "step": 2980
    },
    {
      "epoch": 0.8129418162044589,
      "grad_norm": 1.9375,
      "learning_rate": 4.427832087406556e-05,
      "loss": 0.1886,
      "step": 2990
    },
    {
      "epoch": 0.8156606851549756,
      "grad_norm": 1.296875,
      "learning_rate": 4.4249568717653824e-05,
      "loss": 0.1383,
      "step": 3000
    },
    {
      "epoch": 0.8156606851549756,
      "eval_loss": 0.14339062571525574,
      "eval_runtime": 100.6893,
      "eval_samples_per_second": 4.966,
      "eval_steps_per_second": 0.159,
      "step": 3000
    },
    {
      "epoch": 0.8183795541054921,
      "grad_norm": 1.6328125,
      "learning_rate": 4.42208165612421e-05,
      "loss": 0.135,
      "step": 3010
    },
    {
      "epoch": 0.8210984230560087,
      "grad_norm": 1.1953125,
      "learning_rate": 4.4192064404830366e-05,
      "loss": 0.1488,
      "step": 3020
    },
    {
      "epoch": 0.8238172920065253,
      "grad_norm": 0.73828125,
      "learning_rate": 4.416331224841863e-05,
      "loss": 0.1769,
      "step": 3030
    },
    {
      "epoch": 0.8265361609570419,
      "grad_norm": 1.328125,
      "learning_rate": 4.413456009200691e-05,
      "loss": 0.1685,
      "step": 3040
    },
    {
      "epoch": 0.8292550299075585,
      "grad_norm": 2.921875,
      "learning_rate": 4.4105807935595175e-05,
      "loss": 0.1939,
      "step": 3050
    },
    {
      "epoch": 0.8292550299075585,
      "eval_loss": 0.15489062666893005,
      "eval_runtime": 101.4894,
      "eval_samples_per_second": 4.927,
      "eval_steps_per_second": 0.158,
      "step": 3050
    },
    {
      "epoch": 0.831973898858075,
      "grad_norm": 2.890625,
      "learning_rate": 4.407705577918344e-05,
      "loss": 0.2003,
      "step": 3060
    },
    {
      "epoch": 0.8346927678085916,
      "grad_norm": 1.0390625,
      "learning_rate": 4.404830362277171e-05,
      "loss": 0.1366,
      "step": 3070
    },
    {
      "epoch": 0.8374116367591082,
      "grad_norm": 2.390625,
      "learning_rate": 4.401955146635998e-05,
      "loss": 0.1828,
      "step": 3080
    },
    {
      "epoch": 0.8401305057096248,
      "grad_norm": 1.4453125,
      "learning_rate": 4.399079930994825e-05,
      "loss": 0.1296,
      "step": 3090
    },
    {
      "epoch": 0.8428493746601414,
      "grad_norm": 1.8359375,
      "learning_rate": 4.396204715353652e-05,
      "loss": 0.1239,
      "step": 3100
    },
    {
      "epoch": 0.8428493746601414,
      "eval_loss": 0.14643749594688416,
      "eval_runtime": 111.2646,
      "eval_samples_per_second": 4.494,
      "eval_steps_per_second": 0.144,
      "step": 3100
    },
    {
      "epoch": 0.845568243610658,
      "grad_norm": 1.4140625,
      "learning_rate": 4.3933294997124785e-05,
      "loss": 0.1723,
      "step": 3110
    },
    {
      "epoch": 0.8482871125611745,
      "grad_norm": 1.3515625,
      "learning_rate": 4.390454284071305e-05,
      "loss": 0.1346,
      "step": 3120
    },
    {
      "epoch": 0.8510059815116912,
      "grad_norm": 1.4609375,
      "learning_rate": 4.387579068430133e-05,
      "loss": 0.1286,
      "step": 3130
    },
    {
      "epoch": 0.8537248504622077,
      "grad_norm": 1.3828125,
      "learning_rate": 4.3847038527889594e-05,
      "loss": 0.1481,
      "step": 3140
    },
    {
      "epoch": 0.8564437194127243,
      "grad_norm": 1.875,
      "learning_rate": 4.381828637147786e-05,
      "loss": 0.1904,
      "step": 3150
    },
    {
      "epoch": 0.8564437194127243,
      "eval_loss": 0.14157812297344208,
      "eval_runtime": 108.1,
      "eval_samples_per_second": 4.625,
      "eval_steps_per_second": 0.148,
      "step": 3150
    },
    {
      "epoch": 0.8591625883632409,
      "grad_norm": 2.171875,
      "learning_rate": 4.378953421506613e-05,
      "loss": 0.1712,
      "step": 3160
    },
    {
      "epoch": 0.8618814573137574,
      "grad_norm": 2.234375,
      "learning_rate": 4.37607820586544e-05,
      "loss": 0.1401,
      "step": 3170
    },
    {
      "epoch": 0.8646003262642741,
      "grad_norm": 2.65625,
      "learning_rate": 4.373202990224267e-05,
      "loss": 0.1374,
      "step": 3180
    },
    {
      "epoch": 0.8673191952147906,
      "grad_norm": 2.671875,
      "learning_rate": 4.370327774583094e-05,
      "loss": 0.1622,
      "step": 3190
    },
    {
      "epoch": 0.8700380641653073,
      "grad_norm": 2.421875,
      "learning_rate": 4.3674525589419205e-05,
      "loss": 0.1599,
      "step": 3200
    },
    {
      "epoch": 0.8700380641653073,
      "eval_loss": 0.14441406726837158,
      "eval_runtime": 100.9188,
      "eval_samples_per_second": 4.954,
      "eval_steps_per_second": 0.159,
      "step": 3200
    },
    {
      "epoch": 0.8727569331158238,
      "grad_norm": 2.21875,
      "learning_rate": 4.364577343300747e-05,
      "loss": 0.1574,
      "step": 3210
    },
    {
      "epoch": 0.8754758020663403,
      "grad_norm": 1.7890625,
      "learning_rate": 4.3617021276595746e-05,
      "loss": 0.17,
      "step": 3220
    },
    {
      "epoch": 0.878194671016857,
      "grad_norm": 0.80859375,
      "learning_rate": 4.358826912018401e-05,
      "loss": 0.1319,
      "step": 3230
    },
    {
      "epoch": 0.8809135399673735,
      "grad_norm": 3.625,
      "learning_rate": 4.355951696377229e-05,
      "loss": 0.1521,
      "step": 3240
    },
    {
      "epoch": 0.8836324089178902,
      "grad_norm": 2.59375,
      "learning_rate": 4.3530764807360555e-05,
      "loss": 0.1301,
      "step": 3250
    },
    {
      "epoch": 0.8836324089178902,
      "eval_loss": 0.1458124965429306,
      "eval_runtime": 102.9309,
      "eval_samples_per_second": 4.858,
      "eval_steps_per_second": 0.155,
      "step": 3250
    },
    {
      "epoch": 0.8863512778684067,
      "grad_norm": 3.109375,
      "learning_rate": 4.350201265094883e-05,
      "loss": 0.1503,
      "step": 3260
    },
    {
      "epoch": 0.8890701468189234,
      "grad_norm": 1.21875,
      "learning_rate": 4.3473260494537096e-05,
      "loss": 0.14,
      "step": 3270
    },
    {
      "epoch": 0.8917890157694399,
      "grad_norm": 1.640625,
      "learning_rate": 4.3444508338125363e-05,
      "loss": 0.155,
      "step": 3280
    },
    {
      "epoch": 0.8945078847199565,
      "grad_norm": 1.90625,
      "learning_rate": 4.341575618171363e-05,
      "loss": 0.1594,
      "step": 3290
    },
    {
      "epoch": 0.8972267536704731,
      "grad_norm": 1.734375,
      "learning_rate": 4.33870040253019e-05,
      "loss": 0.1854,
      "step": 3300
    },
    {
      "epoch": 0.8972267536704731,
      "eval_loss": 0.13501562178134918,
      "eval_runtime": 102.9527,
      "eval_samples_per_second": 4.857,
      "eval_steps_per_second": 0.155,
      "step": 3300
    },
    {
      "epoch": 0.8999456226209896,
      "grad_norm": 1.0078125,
      "learning_rate": 4.335825186889017e-05,
      "loss": 0.1385,
      "step": 3310
    },
    {
      "epoch": 0.9026644915715063,
      "grad_norm": 2.171875,
      "learning_rate": 4.332949971247844e-05,
      "loss": 0.1495,
      "step": 3320
    },
    {
      "epoch": 0.9053833605220228,
      "grad_norm": 2.21875,
      "learning_rate": 4.330074755606671e-05,
      "loss": 0.1644,
      "step": 3330
    },
    {
      "epoch": 0.9081022294725394,
      "grad_norm": 4.09375,
      "learning_rate": 4.3271995399654974e-05,
      "loss": 0.1814,
      "step": 3340
    },
    {
      "epoch": 0.910821098423056,
      "grad_norm": 3.78125,
      "learning_rate": 4.324324324324325e-05,
      "loss": 0.1478,
      "step": 3350
    },
    {
      "epoch": 0.910821098423056,
      "eval_loss": 0.1481953114271164,
      "eval_runtime": 127.4811,
      "eval_samples_per_second": 3.922,
      "eval_steps_per_second": 0.126,
      "step": 3350
    },
    {
      "epoch": 0.9135399673735726,
      "grad_norm": 2.125,
      "learning_rate": 4.3214491086831516e-05,
      "loss": 0.1408,
      "step": 3360
    },
    {
      "epoch": 0.9162588363240892,
      "grad_norm": 2.453125,
      "learning_rate": 4.318573893041978e-05,
      "loss": 0.1397,
      "step": 3370
    },
    {
      "epoch": 0.9189777052746058,
      "grad_norm": 1.515625,
      "learning_rate": 4.315698677400805e-05,
      "loss": 0.1547,
      "step": 3380
    },
    {
      "epoch": 0.9216965742251223,
      "grad_norm": 1.5078125,
      "learning_rate": 4.312823461759632e-05,
      "loss": 0.1404,
      "step": 3390
    },
    {
      "epoch": 0.924415443175639,
      "grad_norm": 1.3828125,
      "learning_rate": 4.309948246118459e-05,
      "loss": 0.165,
      "step": 3400
    },
    {
      "epoch": 0.924415443175639,
      "eval_loss": 0.14045313000679016,
      "eval_runtime": 122.3064,
      "eval_samples_per_second": 4.088,
      "eval_steps_per_second": 0.131,
      "step": 3400
    },
    {
      "epoch": 0.9271343121261555,
      "grad_norm": 1.9921875,
      "learning_rate": 4.307073030477286e-05,
      "loss": 0.1744,
      "step": 3410
    },
    {
      "epoch": 0.9298531810766721,
      "grad_norm": 1.6015625,
      "learning_rate": 4.3041978148361126e-05,
      "loss": 0.1302,
      "step": 3420
    },
    {
      "epoch": 0.9325720500271887,
      "grad_norm": 2.046875,
      "learning_rate": 4.3013225991949394e-05,
      "loss": 0.1511,
      "step": 3430
    },
    {
      "epoch": 0.9352909189777052,
      "grad_norm": 2.484375,
      "learning_rate": 4.298447383553767e-05,
      "loss": 0.1756,
      "step": 3440
    },
    {
      "epoch": 0.9380097879282219,
      "grad_norm": 2.109375,
      "learning_rate": 4.2955721679125935e-05,
      "loss": 0.1273,
      "step": 3450
    },
    {
      "epoch": 0.9380097879282219,
      "eval_loss": 0.13562500476837158,
      "eval_runtime": 175.9118,
      "eval_samples_per_second": 2.842,
      "eval_steps_per_second": 0.091,
      "step": 3450
    },
    {
      "epoch": 0.9407286568787384,
      "grad_norm": 1.1171875,
      "learning_rate": 4.29269695227142e-05,
      "loss": 0.1709,
      "step": 3460
    },
    {
      "epoch": 0.9434475258292551,
      "grad_norm": 1.5546875,
      "learning_rate": 4.2898217366302476e-05,
      "loss": 0.1762,
      "step": 3470
    },
    {
      "epoch": 0.9461663947797716,
      "grad_norm": 4.1875,
      "learning_rate": 4.2869465209890744e-05,
      "loss": 0.1788,
      "step": 3480
    },
    {
      "epoch": 0.9488852637302883,
      "grad_norm": 1.5625,
      "learning_rate": 4.284071305347902e-05,
      "loss": 0.1882,
      "step": 3490
    },
    {
      "epoch": 0.9516041326808048,
      "grad_norm": 1.171875,
      "learning_rate": 4.2811960897067285e-05,
      "loss": 0.1873,
      "step": 3500
    },
    {
      "epoch": 0.9516041326808048,
      "eval_loss": 0.14018750190734863,
      "eval_runtime": 130.0961,
      "eval_samples_per_second": 3.843,
      "eval_steps_per_second": 0.123,
      "step": 3500
    },
    {
      "epoch": 0.9543230016313213,
      "grad_norm": 2.359375,
      "learning_rate": 4.278320874065555e-05,
      "loss": 0.1676,
      "step": 3510
    },
    {
      "epoch": 0.957041870581838,
      "grad_norm": 1.4375,
      "learning_rate": 4.275445658424382e-05,
      "loss": 0.1678,
      "step": 3520
    },
    {
      "epoch": 0.9597607395323545,
      "grad_norm": 1.9765625,
      "learning_rate": 4.2725704427832094e-05,
      "loss": 0.1549,
      "step": 3530
    },
    {
      "epoch": 0.9624796084828712,
      "grad_norm": 1.484375,
      "learning_rate": 4.269695227142036e-05,
      "loss": 0.1403,
      "step": 3540
    },
    {
      "epoch": 0.9651984774333877,
      "grad_norm": 2.609375,
      "learning_rate": 4.266820011500863e-05,
      "loss": 0.1458,
      "step": 3550
    },
    {
      "epoch": 0.9651984774333877,
      "eval_loss": 0.13320311903953552,
      "eval_runtime": 100.6993,
      "eval_samples_per_second": 4.965,
      "eval_steps_per_second": 0.159,
      "step": 3550
    },
    {
      "epoch": 0.9679173463839043,
      "grad_norm": 5.46875,
      "learning_rate": 4.2639447958596896e-05,
      "loss": 0.1617,
      "step": 3560
    },
    {
      "epoch": 0.9706362153344209,
      "grad_norm": 2.421875,
      "learning_rate": 4.261069580218516e-05,
      "loss": 0.1612,
      "step": 3570
    },
    {
      "epoch": 0.9733550842849374,
      "grad_norm": 1.421875,
      "learning_rate": 4.258194364577344e-05,
      "loss": 0.1429,
      "step": 3580
    },
    {
      "epoch": 0.9760739532354541,
      "grad_norm": 2.734375,
      "learning_rate": 4.2553191489361704e-05,
      "loss": 0.1602,
      "step": 3590
    },
    {
      "epoch": 0.9787928221859706,
      "grad_norm": 1.4453125,
      "learning_rate": 4.252443933294997e-05,
      "loss": 0.1615,
      "step": 3600
    },
    {
      "epoch": 0.9787928221859706,
      "eval_loss": 0.15165625512599945,
      "eval_runtime": 141.2028,
      "eval_samples_per_second": 3.541,
      "eval_steps_per_second": 0.113,
      "step": 3600
    },
    {
      "epoch": 0.9815116911364872,
      "grad_norm": 1.015625,
      "learning_rate": 4.249568717653824e-05,
      "loss": 0.1485,
      "step": 3610
    },
    {
      "epoch": 0.9842305600870038,
      "grad_norm": 1.546875,
      "learning_rate": 4.246693502012651e-05,
      "loss": 0.1417,
      "step": 3620
    },
    {
      "epoch": 0.9869494290375204,
      "grad_norm": 0.88671875,
      "learning_rate": 4.243818286371478e-05,
      "loss": 0.1246,
      "step": 3630
    },
    {
      "epoch": 0.989668297988037,
      "grad_norm": 0.890625,
      "learning_rate": 4.240943070730305e-05,
      "loss": 0.1586,
      "step": 3640
    },
    {
      "epoch": 0.9923871669385536,
      "grad_norm": 1.234375,
      "learning_rate": 4.2380678550891315e-05,
      "loss": 0.132,
      "step": 3650
    },
    {
      "epoch": 0.9923871669385536,
      "eval_loss": 0.132750004529953,
      "eval_runtime": 103.1912,
      "eval_samples_per_second": 4.845,
      "eval_steps_per_second": 0.155,
      "step": 3650
    },
    {
      "epoch": 0.9951060358890701,
      "grad_norm": 1.9296875,
      "learning_rate": 4.235192639447959e-05,
      "loss": 0.166,
      "step": 3660
    },
    {
      "epoch": 0.9978249048395867,
      "grad_norm": 2.109375,
      "learning_rate": 4.2323174238067857e-05,
      "loss": 0.1289,
      "step": 3670
    },
    {
      "epoch": 1.0005437737901033,
      "grad_norm": 1.453125,
      "learning_rate": 4.2294422081656124e-05,
      "loss": 0.1783,
      "step": 3680
    },
    {
      "epoch": 1.00326264274062,
      "grad_norm": 1.03125,
      "learning_rate": 4.226566992524439e-05,
      "loss": 0.117,
      "step": 3690
    },
    {
      "epoch": 1.0059815116911366,
      "grad_norm": 1.265625,
      "learning_rate": 4.2236917768832665e-05,
      "loss": 0.1481,
      "step": 3700
    },
    {
      "epoch": 1.0059815116911366,
      "eval_loss": 0.1254609376192093,
      "eval_runtime": 98.6559,
      "eval_samples_per_second": 5.068,
      "eval_steps_per_second": 0.162,
      "step": 3700
    },
    {
      "epoch": 1.008700380641653,
      "grad_norm": 1.109375,
      "learning_rate": 4.220816561242093e-05,
      "loss": 0.1283,
      "step": 3710
    },
    {
      "epoch": 1.0114192495921697,
      "grad_norm": 1.65625,
      "learning_rate": 4.217941345600921e-05,
      "loss": 0.1332,
      "step": 3720
    },
    {
      "epoch": 1.0141381185426863,
      "grad_norm": 1.28125,
      "learning_rate": 4.2150661299597474e-05,
      "loss": 0.1439,
      "step": 3730
    },
    {
      "epoch": 1.0168569874932027,
      "grad_norm": 1.453125,
      "learning_rate": 4.212190914318574e-05,
      "loss": 0.1297,
      "step": 3740
    },
    {
      "epoch": 1.0195758564437194,
      "grad_norm": 2.78125,
      "learning_rate": 4.2093156986774015e-05,
      "loss": 0.1396,
      "step": 3750
    },
    {
      "epoch": 1.0195758564437194,
      "eval_loss": 0.13164062798023224,
      "eval_runtime": 98.0163,
      "eval_samples_per_second": 5.101,
      "eval_steps_per_second": 0.163,
      "step": 3750
    },
    {
      "epoch": 1.022294725394236,
      "grad_norm": 1.8046875,
      "learning_rate": 4.206440483036228e-05,
      "loss": 0.1448,
      "step": 3760
    },
    {
      "epoch": 1.0250135943447525,
      "grad_norm": 1.5,
      "learning_rate": 4.203565267395055e-05,
      "loss": 0.1001,
      "step": 3770
    },
    {
      "epoch": 1.0277324632952691,
      "grad_norm": 2.71875,
      "learning_rate": 4.200690051753882e-05,
      "loss": 0.1326,
      "step": 3780
    },
    {
      "epoch": 1.0304513322457858,
      "grad_norm": 1.2578125,
      "learning_rate": 4.1978148361127085e-05,
      "loss": 0.1291,
      "step": 3790
    },
    {
      "epoch": 1.0331702011963024,
      "grad_norm": 1.9140625,
      "learning_rate": 4.194939620471536e-05,
      "loss": 0.15,
      "step": 3800
    },
    {
      "epoch": 1.0331702011963024,
      "eval_loss": 0.1332734376192093,
      "eval_runtime": 103.4538,
      "eval_samples_per_second": 4.833,
      "eval_steps_per_second": 0.155,
      "step": 3800
    },
    {
      "epoch": 1.0358890701468189,
      "grad_norm": 2.3125,
      "learning_rate": 4.1920644048303626e-05,
      "loss": 0.1464,
      "step": 3810
    },
    {
      "epoch": 1.0386079390973355,
      "grad_norm": 1.6015625,
      "learning_rate": 4.189189189189189e-05,
      "loss": 0.1196,
      "step": 3820
    },
    {
      "epoch": 1.0413268080478522,
      "grad_norm": 1.9375,
      "learning_rate": 4.186313973548016e-05,
      "loss": 0.1443,
      "step": 3830
    },
    {
      "epoch": 1.0440456769983686,
      "grad_norm": 1.515625,
      "learning_rate": 4.1834387579068435e-05,
      "loss": 0.135,
      "step": 3840
    },
    {
      "epoch": 1.0467645459488852,
      "grad_norm": 1.3984375,
      "learning_rate": 4.18056354226567e-05,
      "loss": 0.1545,
      "step": 3850
    },
    {
      "epoch": 1.0467645459488852,
      "eval_loss": 0.1333593726158142,
      "eval_runtime": 112.9672,
      "eval_samples_per_second": 4.426,
      "eval_steps_per_second": 0.142,
      "step": 3850
    },
    {
      "epoch": 1.049483414899402,
      "grad_norm": 1.5859375,
      "learning_rate": 4.177688326624497e-05,
      "loss": 0.1351,
      "step": 3860
    },
    {
      "epoch": 1.0522022838499185,
      "grad_norm": 3.578125,
      "learning_rate": 4.174813110983324e-05,
      "loss": 0.1514,
      "step": 3870
    },
    {
      "epoch": 1.054921152800435,
      "grad_norm": 1.421875,
      "learning_rate": 4.1719378953421504e-05,
      "loss": 0.1232,
      "step": 3880
    },
    {
      "epoch": 1.0576400217509516,
      "grad_norm": 2.71875,
      "learning_rate": 4.169062679700978e-05,
      "loss": 0.1487,
      "step": 3890
    },
    {
      "epoch": 1.0603588907014683,
      "grad_norm": 2.078125,
      "learning_rate": 4.1661874640598045e-05,
      "loss": 0.1357,
      "step": 3900
    },
    {
      "epoch": 1.0603588907014683,
      "eval_loss": 0.12790624797344208,
      "eval_runtime": 104.8343,
      "eval_samples_per_second": 4.769,
      "eval_steps_per_second": 0.153,
      "step": 3900
    },
    {
      "epoch": 1.0630777596519847,
      "grad_norm": 1.8828125,
      "learning_rate": 4.163312248418631e-05,
      "loss": 0.141,
      "step": 3910
    },
    {
      "epoch": 1.0657966286025014,
      "grad_norm": 1.96875,
      "learning_rate": 4.160437032777458e-05,
      "loss": 0.1453,
      "step": 3920
    },
    {
      "epoch": 1.068515497553018,
      "grad_norm": 2.0625,
      "learning_rate": 4.1575618171362854e-05,
      "loss": 0.1122,
      "step": 3930
    },
    {
      "epoch": 1.0712343665035344,
      "grad_norm": 2.203125,
      "learning_rate": 4.154686601495112e-05,
      "loss": 0.1314,
      "step": 3940
    },
    {
      "epoch": 1.073953235454051,
      "grad_norm": 1.2734375,
      "learning_rate": 4.1518113858539396e-05,
      "loss": 0.1253,
      "step": 3950
    },
    {
      "epoch": 1.073953235454051,
      "eval_loss": 0.1352812498807907,
      "eval_runtime": 104.9923,
      "eval_samples_per_second": 4.762,
      "eval_steps_per_second": 0.152,
      "step": 3950
    },
    {
      "epoch": 1.0766721044045677,
      "grad_norm": 2.5,
      "learning_rate": 4.148936170212766e-05,
      "loss": 0.1598,
      "step": 3960
    },
    {
      "epoch": 1.0793909733550844,
      "grad_norm": 1.953125,
      "learning_rate": 4.146060954571593e-05,
      "loss": 0.127,
      "step": 3970
    },
    {
      "epoch": 1.0821098423056008,
      "grad_norm": 1.875,
      "learning_rate": 4.1431857389304204e-05,
      "loss": 0.1478,
      "step": 3980
    },
    {
      "epoch": 1.0848287112561175,
      "grad_norm": 0.6328125,
      "learning_rate": 4.140310523289247e-05,
      "loss": 0.1445,
      "step": 3990
    },
    {
      "epoch": 1.0875475802066341,
      "grad_norm": 1.09375,
      "learning_rate": 4.137435307648074e-05,
      "loss": 0.1509,
      "step": 4000
    },
    {
      "epoch": 1.0875475802066341,
      "eval_loss": 0.14360156655311584,
      "eval_runtime": 103.6969,
      "eval_samples_per_second": 4.822,
      "eval_steps_per_second": 0.154,
      "step": 4000
    },
    {
      "epoch": 1.0902664491571505,
      "grad_norm": 1.25,
      "learning_rate": 4.1345600920069006e-05,
      "loss": 0.1226,
      "step": 4010
    },
    {
      "epoch": 1.0929853181076672,
      "grad_norm": 2.09375,
      "learning_rate": 4.131684876365728e-05,
      "loss": 0.1265,
      "step": 4020
    },
    {
      "epoch": 1.0957041870581838,
      "grad_norm": 2.140625,
      "learning_rate": 4.128809660724555e-05,
      "loss": 0.1417,
      "step": 4030
    },
    {
      "epoch": 1.0984230560087003,
      "grad_norm": 1.328125,
      "learning_rate": 4.1259344450833815e-05,
      "loss": 0.1393,
      "step": 4040
    },
    {
      "epoch": 1.101141924959217,
      "grad_norm": 3.46875,
      "learning_rate": 4.123059229442208e-05,
      "loss": 0.127,
      "step": 4050
    },
    {
      "epoch": 1.101141924959217,
      "eval_loss": 0.14453125,
      "eval_runtime": 103.7137,
      "eval_samples_per_second": 4.821,
      "eval_steps_per_second": 0.154,
      "step": 4050
    },
    {
      "epoch": 1.1038607939097336,
      "grad_norm": 1.9765625,
      "learning_rate": 4.120184013801035e-05,
      "loss": 0.1353,
      "step": 4060
    },
    {
      "epoch": 1.1065796628602502,
      "grad_norm": 1.390625,
      "learning_rate": 4.1173087981598624e-05,
      "loss": 0.1293,
      "step": 4070
    },
    {
      "epoch": 1.1092985318107667,
      "grad_norm": 1.921875,
      "learning_rate": 4.114433582518689e-05,
      "loss": 0.143,
      "step": 4080
    },
    {
      "epoch": 1.1120174007612833,
      "grad_norm": 2.609375,
      "learning_rate": 4.111558366877516e-05,
      "loss": 0.1391,
      "step": 4090
    },
    {
      "epoch": 1.1147362697118,
      "grad_norm": 2.59375,
      "learning_rate": 4.1086831512363426e-05,
      "loss": 0.1398,
      "step": 4100
    },
    {
      "epoch": 1.1147362697118,
      "eval_loss": 0.14257031679153442,
      "eval_runtime": 102.7828,
      "eval_samples_per_second": 4.865,
      "eval_steps_per_second": 0.156,
      "step": 4100
    },
    {
      "epoch": 1.1174551386623164,
      "grad_norm": 2.515625,
      "learning_rate": 4.10580793559517e-05,
      "loss": 0.1375,
      "step": 4110
    },
    {
      "epoch": 1.120174007612833,
      "grad_norm": 0.703125,
      "learning_rate": 4.102932719953997e-05,
      "loss": 0.1563,
      "step": 4120
    },
    {
      "epoch": 1.1228928765633497,
      "grad_norm": 1.359375,
      "learning_rate": 4.1000575043128234e-05,
      "loss": 0.1312,
      "step": 4130
    },
    {
      "epoch": 1.1256117455138663,
      "grad_norm": 2.25,
      "learning_rate": 4.09718228867165e-05,
      "loss": 0.1371,
      "step": 4140
    },
    {
      "epoch": 1.1283306144643828,
      "grad_norm": 3.359375,
      "learning_rate": 4.094307073030477e-05,
      "loss": 0.1291,
      "step": 4150
    },
    {
      "epoch": 1.1283306144643828,
      "eval_loss": 0.14207030832767487,
      "eval_runtime": 102.7106,
      "eval_samples_per_second": 4.868,
      "eval_steps_per_second": 0.156,
      "step": 4150
    },
    {
      "epoch": 1.1310494834148994,
      "grad_norm": 4.375,
      "learning_rate": 4.091431857389304e-05,
      "loss": 0.1438,
      "step": 4160
    },
    {
      "epoch": 1.133768352365416,
      "grad_norm": 2.375,
      "learning_rate": 4.088556641748131e-05,
      "loss": 0.1373,
      "step": 4170
    },
    {
      "epoch": 1.1364872213159325,
      "grad_norm": 3.203125,
      "learning_rate": 4.0856814261069584e-05,
      "loss": 0.1513,
      "step": 4180
    },
    {
      "epoch": 1.1392060902664491,
      "grad_norm": 1.2734375,
      "learning_rate": 4.082806210465785e-05,
      "loss": 0.1152,
      "step": 4190
    },
    {
      "epoch": 1.1419249592169658,
      "grad_norm": 4.03125,
      "learning_rate": 4.0799309948246126e-05,
      "loss": 0.1606,
      "step": 4200
    },
    {
      "epoch": 1.1419249592169658,
      "eval_loss": 0.13499218225479126,
      "eval_runtime": 106.1654,
      "eval_samples_per_second": 4.71,
      "eval_steps_per_second": 0.151,
      "step": 4200
    },
    {
      "epoch": 1.1446438281674824,
      "grad_norm": 1.5625,
      "learning_rate": 4.077055779183439e-05,
      "loss": 0.1257,
      "step": 4210
    },
    {
      "epoch": 1.1473626971179989,
      "grad_norm": 1.1171875,
      "learning_rate": 4.074180563542266e-05,
      "loss": 0.1324,
      "step": 4220
    },
    {
      "epoch": 1.1500815660685155,
      "grad_norm": 0.90625,
      "learning_rate": 4.071305347901093e-05,
      "loss": 0.1352,
      "step": 4230
    },
    {
      "epoch": 1.1528004350190322,
      "grad_norm": 2.09375,
      "learning_rate": 4.0684301322599195e-05,
      "loss": 0.1353,
      "step": 4240
    },
    {
      "epoch": 1.1555193039695486,
      "grad_norm": 2.078125,
      "learning_rate": 4.065554916618747e-05,
      "loss": 0.1319,
      "step": 4250
    },
    {
      "epoch": 1.1555193039695486,
      "eval_loss": 0.12882812321186066,
      "eval_runtime": 102.2447,
      "eval_samples_per_second": 4.89,
      "eval_steps_per_second": 0.156,
      "step": 4250
    },
    {
      "epoch": 1.1582381729200653,
      "grad_norm": 2.078125,
      "learning_rate": 4.0626797009775737e-05,
      "loss": 0.1521,
      "step": 4260
    },
    {
      "epoch": 1.160957041870582,
      "grad_norm": 2.046875,
      "learning_rate": 4.0598044853364004e-05,
      "loss": 0.1399,
      "step": 4270
    },
    {
      "epoch": 1.1636759108210983,
      "grad_norm": 1.9140625,
      "learning_rate": 4.056929269695227e-05,
      "loss": 0.1477,
      "step": 4280
    },
    {
      "epoch": 1.166394779771615,
      "grad_norm": 2.21875,
      "learning_rate": 4.0540540540540545e-05,
      "loss": 0.1498,
      "step": 4290
    },
    {
      "epoch": 1.1691136487221316,
      "grad_norm": 1.53125,
      "learning_rate": 4.051178838412881e-05,
      "loss": 0.1309,
      "step": 4300
    },
    {
      "epoch": 1.1691136487221316,
      "eval_loss": 0.13154686987400055,
      "eval_runtime": 108.064,
      "eval_samples_per_second": 4.627,
      "eval_steps_per_second": 0.148,
      "step": 4300
    },
    {
      "epoch": 1.171832517672648,
      "grad_norm": 2.203125,
      "learning_rate": 4.048303622771708e-05,
      "loss": 0.1411,
      "step": 4310
    },
    {
      "epoch": 1.1745513866231647,
      "grad_norm": 1.25,
      "learning_rate": 4.045428407130535e-05,
      "loss": 0.1168,
      "step": 4320
    },
    {
      "epoch": 1.1772702555736814,
      "grad_norm": 2.578125,
      "learning_rate": 4.0425531914893614e-05,
      "loss": 0.123,
      "step": 4330
    },
    {
      "epoch": 1.179989124524198,
      "grad_norm": 1.5,
      "learning_rate": 4.039677975848189e-05,
      "loss": 0.1097,
      "step": 4340
    },
    {
      "epoch": 1.1827079934747144,
      "grad_norm": 1.34375,
      "learning_rate": 4.0368027602070156e-05,
      "loss": 0.1183,
      "step": 4350
    },
    {
      "epoch": 1.1827079934747144,
      "eval_loss": 0.1254609376192093,
      "eval_runtime": 116.5666,
      "eval_samples_per_second": 4.289,
      "eval_steps_per_second": 0.137,
      "step": 4350
    },
    {
      "epoch": 1.185426862425231,
      "grad_norm": 1.4375,
      "learning_rate": 4.033927544565842e-05,
      "loss": 0.1022,
      "step": 4360
    },
    {
      "epoch": 1.1881457313757477,
      "grad_norm": 1.5859375,
      "learning_rate": 4.031052328924669e-05,
      "loss": 0.1182,
      "step": 4370
    },
    {
      "epoch": 1.1908646003262642,
      "grad_norm": 2.109375,
      "learning_rate": 4.0281771132834965e-05,
      "loss": 0.1232,
      "step": 4380
    },
    {
      "epoch": 1.1935834692767808,
      "grad_norm": 2.21875,
      "learning_rate": 4.025301897642323e-05,
      "loss": 0.1069,
      "step": 4390
    },
    {
      "epoch": 1.1963023382272975,
      "grad_norm": 1.0546875,
      "learning_rate": 4.02242668200115e-05,
      "loss": 0.1869,
      "step": 4400
    },
    {
      "epoch": 1.1963023382272975,
      "eval_loss": 0.13631249964237213,
      "eval_runtime": 102.6603,
      "eval_samples_per_second": 4.87,
      "eval_steps_per_second": 0.156,
      "step": 4400
    },
    {
      "epoch": 1.1990212071778141,
      "grad_norm": 2.578125,
      "learning_rate": 4.019551466359977e-05,
      "loss": 0.1576,
      "step": 4410
    },
    {
      "epoch": 1.2017400761283306,
      "grad_norm": 0.97265625,
      "learning_rate": 4.016676250718804e-05,
      "loss": 0.134,
      "step": 4420
    },
    {
      "epoch": 1.2044589450788472,
      "grad_norm": 1.125,
      "learning_rate": 4.0138010350776315e-05,
      "loss": 0.1528,
      "step": 4430
    },
    {
      "epoch": 1.2071778140293639,
      "grad_norm": 8.9375,
      "learning_rate": 4.010925819436458e-05,
      "loss": 0.1315,
      "step": 4440
    },
    {
      "epoch": 1.2098966829798803,
      "grad_norm": 2.03125,
      "learning_rate": 4.008050603795285e-05,
      "loss": 0.1478,
      "step": 4450
    },
    {
      "epoch": 1.2098966829798803,
      "eval_loss": 0.12761718034744263,
      "eval_runtime": 103.2637,
      "eval_samples_per_second": 4.842,
      "eval_steps_per_second": 0.155,
      "step": 4450
    },
    {
      "epoch": 1.212615551930397,
      "grad_norm": 1.6015625,
      "learning_rate": 4.005175388154112e-05,
      "loss": 0.131,
      "step": 4460
    },
    {
      "epoch": 1.2153344208809136,
      "grad_norm": 2.28125,
      "learning_rate": 4.002300172512939e-05,
      "loss": 0.0993,
      "step": 4470
    },
    {
      "epoch": 1.2180532898314302,
      "grad_norm": 1.859375,
      "learning_rate": 3.999424956871766e-05,
      "loss": 0.1299,
      "step": 4480
    },
    {
      "epoch": 1.2207721587819467,
      "grad_norm": 2.265625,
      "learning_rate": 3.9965497412305925e-05,
      "loss": 0.144,
      "step": 4490
    },
    {
      "epoch": 1.2234910277324633,
      "grad_norm": 1.3203125,
      "learning_rate": 3.993674525589419e-05,
      "loss": 0.1627,
      "step": 4500
    },
    {
      "epoch": 1.2234910277324633,
      "eval_loss": 0.12840625643730164,
      "eval_runtime": 102.6742,
      "eval_samples_per_second": 4.87,
      "eval_steps_per_second": 0.156,
      "step": 4500
    },
    {
      "epoch": 1.22620989668298,
      "grad_norm": 0.80859375,
      "learning_rate": 3.990799309948247e-05,
      "loss": 0.1348,
      "step": 4510
    },
    {
      "epoch": 1.2289287656334964,
      "grad_norm": 1.421875,
      "learning_rate": 3.9879240943070734e-05,
      "loss": 0.1186,
      "step": 4520
    },
    {
      "epoch": 1.231647634584013,
      "grad_norm": 2.90625,
      "learning_rate": 3.9850488786659e-05,
      "loss": 0.1382,
      "step": 4530
    },
    {
      "epoch": 1.2343665035345297,
      "grad_norm": 1.5234375,
      "learning_rate": 3.982173663024727e-05,
      "loss": 0.1232,
      "step": 4540
    },
    {
      "epoch": 1.2370853724850461,
      "grad_norm": 3.046875,
      "learning_rate": 3.9792984473835536e-05,
      "loss": 0.1675,
      "step": 4550
    },
    {
      "epoch": 1.2370853724850461,
      "eval_loss": 0.12412499636411667,
      "eval_runtime": 102.4426,
      "eval_samples_per_second": 4.881,
      "eval_steps_per_second": 0.156,
      "step": 4550
    },
    {
      "epoch": 1.2398042414355628,
      "grad_norm": 1.203125,
      "learning_rate": 3.976423231742381e-05,
      "loss": 0.1151,
      "step": 4560
    },
    {
      "epoch": 1.2425231103860794,
      "grad_norm": 1.1875,
      "learning_rate": 3.973548016101208e-05,
      "loss": 0.1278,
      "step": 4570
    },
    {
      "epoch": 1.2452419793365959,
      "grad_norm": 1.765625,
      "learning_rate": 3.9706728004600345e-05,
      "loss": 0.129,
      "step": 4580
    },
    {
      "epoch": 1.2479608482871125,
      "grad_norm": 2.46875,
      "learning_rate": 3.967797584818861e-05,
      "loss": 0.1575,
      "step": 4590
    },
    {
      "epoch": 1.2506797172376292,
      "grad_norm": 2.265625,
      "learning_rate": 3.9649223691776886e-05,
      "loss": 0.1228,
      "step": 4600
    },
    {
      "epoch": 1.2506797172376292,
      "eval_loss": 0.13243749737739563,
      "eval_runtime": 102.2112,
      "eval_samples_per_second": 4.892,
      "eval_steps_per_second": 0.157,
      "step": 4600
    },
    {
      "epoch": 1.2533985861881458,
      "grad_norm": 3.125,
      "learning_rate": 3.9620471535365154e-05,
      "loss": 0.1423,
      "step": 4610
    },
    {
      "epoch": 1.2561174551386622,
      "grad_norm": 1.703125,
      "learning_rate": 3.959171937895342e-05,
      "loss": 0.1021,
      "step": 4620
    },
    {
      "epoch": 1.258836324089179,
      "grad_norm": 2.015625,
      "learning_rate": 3.956296722254169e-05,
      "loss": 0.1194,
      "step": 4630
    },
    {
      "epoch": 1.2615551930396955,
      "grad_norm": 1.4765625,
      "learning_rate": 3.953421506612996e-05,
      "loss": 0.111,
      "step": 4640
    },
    {
      "epoch": 1.264274061990212,
      "grad_norm": 2.046875,
      "learning_rate": 3.950546290971823e-05,
      "loss": 0.1131,
      "step": 4650
    },
    {
      "epoch": 1.264274061990212,
      "eval_loss": 0.13734374940395355,
      "eval_runtime": 98.9572,
      "eval_samples_per_second": 5.053,
      "eval_steps_per_second": 0.162,
      "step": 4650
    },
    {
      "epoch": 1.2669929309407286,
      "grad_norm": 1.6875,
      "learning_rate": 3.9476710753306504e-05,
      "loss": 0.1537,
      "step": 4660
    },
    {
      "epoch": 1.2697117998912453,
      "grad_norm": 1.2734375,
      "learning_rate": 3.944795859689477e-05,
      "loss": 0.1523,
      "step": 4670
    },
    {
      "epoch": 1.272430668841762,
      "grad_norm": 2.671875,
      "learning_rate": 3.941920644048304e-05,
      "loss": 0.1143,
      "step": 4680
    },
    {
      "epoch": 1.2751495377922784,
      "grad_norm": 2.3125,
      "learning_rate": 3.939045428407131e-05,
      "loss": 0.1253,
      "step": 4690
    },
    {
      "epoch": 1.277868406742795,
      "grad_norm": 1.671875,
      "learning_rate": 3.936170212765958e-05,
      "loss": 0.1373,
      "step": 4700
    },
    {
      "epoch": 1.277868406742795,
      "eval_loss": 0.1303749978542328,
      "eval_runtime": 102.6121,
      "eval_samples_per_second": 4.873,
      "eval_steps_per_second": 0.156,
      "step": 4700
    },
    {
      "epoch": 1.2805872756933117,
      "grad_norm": 2.328125,
      "learning_rate": 3.933294997124785e-05,
      "loss": 0.1458,
      "step": 4710
    },
    {
      "epoch": 1.283306144643828,
      "grad_norm": 3.390625,
      "learning_rate": 3.9304197814836114e-05,
      "loss": 0.1472,
      "step": 4720
    },
    {
      "epoch": 1.2860250135943447,
      "grad_norm": 1.3671875,
      "learning_rate": 3.927544565842438e-05,
      "loss": 0.1301,
      "step": 4730
    },
    {
      "epoch": 1.2887438825448614,
      "grad_norm": 2.484375,
      "learning_rate": 3.9246693502012656e-05,
      "loss": 0.1159,
      "step": 4740
    },
    {
      "epoch": 1.291462751495378,
      "grad_norm": 1.140625,
      "learning_rate": 3.921794134560092e-05,
      "loss": 0.1414,
      "step": 4750
    },
    {
      "epoch": 1.291462751495378,
      "eval_loss": 0.1350078135728836,
      "eval_runtime": 102.0545,
      "eval_samples_per_second": 4.899,
      "eval_steps_per_second": 0.157,
      "step": 4750
    },
    {
      "epoch": 1.2941816204458945,
      "grad_norm": 1.15625,
      "learning_rate": 3.918918918918919e-05,
      "loss": 0.1459,
      "step": 4760
    },
    {
      "epoch": 1.2969004893964111,
      "grad_norm": 1.171875,
      "learning_rate": 3.916043703277746e-05,
      "loss": 0.1386,
      "step": 4770
    },
    {
      "epoch": 1.2996193583469275,
      "grad_norm": 2.96875,
      "learning_rate": 3.913168487636573e-05,
      "loss": 0.1703,
      "step": 4780
    },
    {
      "epoch": 1.3023382272974442,
      "grad_norm": 1.515625,
      "learning_rate": 3.9102932719954e-05,
      "loss": 0.1592,
      "step": 4790
    },
    {
      "epoch": 1.3050570962479608,
      "grad_norm": 1.8125,
      "learning_rate": 3.9074180563542266e-05,
      "loss": 0.1163,
      "step": 4800
    },
    {
      "epoch": 1.3050570962479608,
      "eval_loss": 0.13824999332427979,
      "eval_runtime": 103.1396,
      "eval_samples_per_second": 4.848,
      "eval_steps_per_second": 0.155,
      "step": 4800
    },
    {
      "epoch": 1.3077759651984775,
      "grad_norm": 1.453125,
      "learning_rate": 3.9045428407130534e-05,
      "loss": 0.137,
      "step": 4810
    },
    {
      "epoch": 1.3104948341489941,
      "grad_norm": 1.0703125,
      "learning_rate": 3.90166762507188e-05,
      "loss": 0.1204,
      "step": 4820
    },
    {
      "epoch": 1.3132137030995106,
      "grad_norm": 1.0859375,
      "learning_rate": 3.8987924094307075e-05,
      "loss": 0.113,
      "step": 4830
    },
    {
      "epoch": 1.3159325720500272,
      "grad_norm": 1.6953125,
      "learning_rate": 3.895917193789534e-05,
      "loss": 0.1437,
      "step": 4840
    },
    {
      "epoch": 1.3186514410005437,
      "grad_norm": 0.80859375,
      "learning_rate": 3.893041978148361e-05,
      "loss": 0.1482,
      "step": 4850
    },
    {
      "epoch": 1.3186514410005437,
      "eval_loss": 0.1264999955892563,
      "eval_runtime": 127.1623,
      "eval_samples_per_second": 3.932,
      "eval_steps_per_second": 0.126,
      "step": 4850
    },
    {
      "epoch": 1.3213703099510603,
      "grad_norm": 0.97265625,
      "learning_rate": 3.890166762507188e-05,
      "loss": 0.1199,
      "step": 4860
    },
    {
      "epoch": 1.324089178901577,
      "grad_norm": 1.3984375,
      "learning_rate": 3.887291546866015e-05,
      "loss": 0.1391,
      "step": 4870
    },
    {
      "epoch": 1.3268080478520936,
      "grad_norm": 1.046875,
      "learning_rate": 3.884416331224842e-05,
      "loss": 0.1064,
      "step": 4880
    },
    {
      "epoch": 1.32952691680261,
      "grad_norm": 2.703125,
      "learning_rate": 3.881541115583669e-05,
      "loss": 0.1202,
      "step": 4890
    },
    {
      "epoch": 1.3322457857531267,
      "grad_norm": 3.203125,
      "learning_rate": 3.878665899942496e-05,
      "loss": 0.1357,
      "step": 4900
    },
    {
      "epoch": 1.3322457857531267,
      "eval_loss": 0.13258594274520874,
      "eval_runtime": 107.6026,
      "eval_samples_per_second": 4.647,
      "eval_steps_per_second": 0.149,
      "step": 4900
    },
    {
      "epoch": 1.3349646547036433,
      "grad_norm": 2.28125,
      "learning_rate": 3.875790684301323e-05,
      "loss": 0.1447,
      "step": 4910
    },
    {
      "epoch": 1.3376835236541598,
      "grad_norm": 2.671875,
      "learning_rate": 3.87291546866015e-05,
      "loss": 0.1394,
      "step": 4920
    },
    {
      "epoch": 1.3404023926046764,
      "grad_norm": 1.6484375,
      "learning_rate": 3.870040253018977e-05,
      "loss": 0.1571,
      "step": 4930
    },
    {
      "epoch": 1.343121261555193,
      "grad_norm": 2.59375,
      "learning_rate": 3.8671650373778036e-05,
      "loss": 0.1277,
      "step": 4940
    },
    {
      "epoch": 1.3458401305057097,
      "grad_norm": 1.0390625,
      "learning_rate": 3.86428982173663e-05,
      "loss": 0.1125,
      "step": 4950
    },
    {
      "epoch": 1.3458401305057097,
      "eval_loss": 0.1306796818971634,
      "eval_runtime": 134.5287,
      "eval_samples_per_second": 3.717,
      "eval_steps_per_second": 0.119,
      "step": 4950
    },
    {
      "epoch": 1.3485589994562261,
      "grad_norm": 1.3671875,
      "learning_rate": 3.861414606095458e-05,
      "loss": 0.1495,
      "step": 4960
    },
    {
      "epoch": 1.3512778684067428,
      "grad_norm": 1.515625,
      "learning_rate": 3.8585393904542845e-05,
      "loss": 0.1405,
      "step": 4970
    },
    {
      "epoch": 1.3539967373572595,
      "grad_norm": 0.8125,
      "learning_rate": 3.855664174813111e-05,
      "loss": 0.1494,
      "step": 4980
    },
    {
      "epoch": 1.3567156063077759,
      "grad_norm": 0.7578125,
      "learning_rate": 3.852788959171938e-05,
      "loss": 0.108,
      "step": 4990
    },
    {
      "epoch": 1.3594344752582925,
      "grad_norm": 2.1875,
      "learning_rate": 3.8499137435307647e-05,
      "loss": 0.1223,
      "step": 5000
    },
    {
      "epoch": 1.3594344752582925,
      "eval_loss": 0.13410937786102295,
      "eval_runtime": 102.63,
      "eval_samples_per_second": 4.872,
      "eval_steps_per_second": 0.156,
      "step": 5000
    },
    {
      "epoch": 1.3621533442088092,
      "grad_norm": 1.2578125,
      "learning_rate": 3.847038527889592e-05,
      "loss": 0.1542,
      "step": 5010
    },
    {
      "epoch": 1.3648722131593258,
      "grad_norm": 1.6484375,
      "learning_rate": 3.844163312248419e-05,
      "loss": 0.1537,
      "step": 5020
    },
    {
      "epoch": 1.3675910821098423,
      "grad_norm": 1.296875,
      "learning_rate": 3.8412880966072455e-05,
      "loss": 0.1329,
      "step": 5030
    },
    {
      "epoch": 1.370309951060359,
      "grad_norm": 1.9609375,
      "learning_rate": 3.838412880966072e-05,
      "loss": 0.1242,
      "step": 5040
    },
    {
      "epoch": 1.3730288200108753,
      "grad_norm": 2.109375,
      "learning_rate": 3.8355376653249e-05,
      "loss": 0.1294,
      "step": 5050
    },
    {
      "epoch": 1.3730288200108753,
      "eval_loss": 0.13114844262599945,
      "eval_runtime": 258.6114,
      "eval_samples_per_second": 1.933,
      "eval_steps_per_second": 0.062,
      "step": 5050
    },
    {
      "epoch": 1.375747688961392,
      "grad_norm": 1.96875,
      "learning_rate": 3.8326624496837264e-05,
      "loss": 0.1438,
      "step": 5060
    },
    {
      "epoch": 1.3784665579119086,
      "grad_norm": 2.296875,
      "learning_rate": 3.829787234042553e-05,
      "loss": 0.1133,
      "step": 5070
    },
    {
      "epoch": 1.3811854268624253,
      "grad_norm": 1.6875,
      "learning_rate": 3.82691201840138e-05,
      "loss": 0.1268,
      "step": 5080
    },
    {
      "epoch": 1.383904295812942,
      "grad_norm": 1.046875,
      "learning_rate": 3.8240368027602066e-05,
      "loss": 0.1173,
      "step": 5090
    },
    {
      "epoch": 1.3866231647634584,
      "grad_norm": 1.0546875,
      "learning_rate": 3.821161587119034e-05,
      "loss": 0.1491,
      "step": 5100
    },
    {
      "epoch": 1.3866231647634584,
      "eval_loss": 0.12962499260902405,
      "eval_runtime": 107.6895,
      "eval_samples_per_second": 4.643,
      "eval_steps_per_second": 0.149,
      "step": 5100
    },
    {
      "epoch": 1.389342033713975,
      "grad_norm": 1.703125,
      "learning_rate": 3.818286371477861e-05,
      "loss": 0.1289,
      "step": 5110
    },
    {
      "epoch": 1.3920609026644915,
      "grad_norm": 1.2421875,
      "learning_rate": 3.815411155836688e-05,
      "loss": 0.1729,
      "step": 5120
    },
    {
      "epoch": 1.394779771615008,
      "grad_norm": 2.265625,
      "learning_rate": 3.812535940195515e-05,
      "loss": 0.1578,
      "step": 5130
    },
    {
      "epoch": 1.3974986405655248,
      "grad_norm": 2.078125,
      "learning_rate": 3.809660724554342e-05,
      "loss": 0.1182,
      "step": 5140
    },
    {
      "epoch": 1.4002175095160414,
      "grad_norm": 0.79296875,
      "learning_rate": 3.806785508913169e-05,
      "loss": 0.1121,
      "step": 5150
    },
    {
      "epoch": 1.4002175095160414,
      "eval_loss": 0.11897656321525574,
      "eval_runtime": 131.3197,
      "eval_samples_per_second": 3.808,
      "eval_steps_per_second": 0.122,
      "step": 5150
    },
    {
      "epoch": 1.4029363784665578,
      "grad_norm": 3.453125,
      "learning_rate": 3.803910293271996e-05,
      "loss": 0.1207,
      "step": 5160
    },
    {
      "epoch": 1.4056552474170745,
      "grad_norm": 2.03125,
      "learning_rate": 3.8010350776308225e-05,
      "loss": 0.1445,
      "step": 5170
    },
    {
      "epoch": 1.4083741163675911,
      "grad_norm": 2.796875,
      "learning_rate": 3.79815986198965e-05,
      "loss": 0.1309,
      "step": 5180
    },
    {
      "epoch": 1.4110929853181076,
      "grad_norm": 1.7265625,
      "learning_rate": 3.7952846463484766e-05,
      "loss": 0.1186,
      "step": 5190
    },
    {
      "epoch": 1.4138118542686242,
      "grad_norm": 1.2734375,
      "learning_rate": 3.7924094307073033e-05,
      "loss": 0.1029,
      "step": 5200
    },
    {
      "epoch": 1.4138118542686242,
      "eval_loss": 0.12324218451976776,
      "eval_runtime": 205.6437,
      "eval_samples_per_second": 2.431,
      "eval_steps_per_second": 0.078,
      "step": 5200
    },
    {
      "epoch": 1.4165307232191409,
      "grad_norm": 1.140625,
      "learning_rate": 3.78953421506613e-05,
      "loss": 0.1234,
      "step": 5210
    },
    {
      "epoch": 1.4192495921696575,
      "grad_norm": 1.6328125,
      "learning_rate": 3.786658999424957e-05,
      "loss": 0.1373,
      "step": 5220
    },
    {
      "epoch": 1.421968461120174,
      "grad_norm": 1.265625,
      "learning_rate": 3.783783783783784e-05,
      "loss": 0.1132,
      "step": 5230
    },
    {
      "epoch": 1.4246873300706906,
      "grad_norm": 1.0703125,
      "learning_rate": 3.780908568142611e-05,
      "loss": 0.0986,
      "step": 5240
    },
    {
      "epoch": 1.4274061990212072,
      "grad_norm": 0.72265625,
      "learning_rate": 3.778033352501438e-05,
      "loss": 0.138,
      "step": 5250
    },
    {
      "epoch": 1.4274061990212072,
      "eval_loss": 0.12297656387090683,
      "eval_runtime": 117.8714,
      "eval_samples_per_second": 4.242,
      "eval_steps_per_second": 0.136,
      "step": 5250
    },
    {
      "epoch": 1.4301250679717237,
      "grad_norm": 2.078125,
      "learning_rate": 3.7751581368602644e-05,
      "loss": 0.1507,
      "step": 5260
    },
    {
      "epoch": 1.4328439369222403,
      "grad_norm": 1.2890625,
      "learning_rate": 3.772282921219092e-05,
      "loss": 0.1283,
      "step": 5270
    },
    {
      "epoch": 1.435562805872757,
      "grad_norm": 1.75,
      "learning_rate": 3.7694077055779186e-05,
      "loss": 0.1409,
      "step": 5280
    },
    {
      "epoch": 1.4382816748232736,
      "grad_norm": 1.0234375,
      "learning_rate": 3.766532489936745e-05,
      "loss": 0.1149,
      "step": 5290
    },
    {
      "epoch": 1.44100054377379,
      "grad_norm": 1.84375,
      "learning_rate": 3.763657274295572e-05,
      "loss": 0.1167,
      "step": 5300
    },
    {
      "epoch": 1.44100054377379,
      "eval_loss": 0.12507031857967377,
      "eval_runtime": 115.6033,
      "eval_samples_per_second": 4.325,
      "eval_steps_per_second": 0.138,
      "step": 5300
    },
    {
      "epoch": 1.4437194127243067,
      "grad_norm": 2.515625,
      "learning_rate": 3.760782058654399e-05,
      "loss": 0.1308,
      "step": 5310
    },
    {
      "epoch": 1.4464382816748234,
      "grad_norm": 1.1875,
      "learning_rate": 3.757906843013226e-05,
      "loss": 0.12,
      "step": 5320
    },
    {
      "epoch": 1.4491571506253398,
      "grad_norm": 1.5390625,
      "learning_rate": 3.755031627372053e-05,
      "loss": 0.1337,
      "step": 5330
    },
    {
      "epoch": 1.4518760195758564,
      "grad_norm": 1.1875,
      "learning_rate": 3.7521564117308796e-05,
      "loss": 0.1345,
      "step": 5340
    },
    {
      "epoch": 1.454594888526373,
      "grad_norm": 0.76171875,
      "learning_rate": 3.749281196089707e-05,
      "loss": 0.116,
      "step": 5350
    },
    {
      "epoch": 1.454594888526373,
      "eval_loss": 0.12296094000339508,
      "eval_runtime": 206.7988,
      "eval_samples_per_second": 2.418,
      "eval_steps_per_second": 0.077,
      "step": 5350
    },
    {
      "epoch": 1.4573137574768897,
      "grad_norm": 1.3125,
      "learning_rate": 3.746405980448534e-05,
      "loss": 0.0951,
      "step": 5360
    },
    {
      "epoch": 1.4600326264274062,
      "grad_norm": 1.6640625,
      "learning_rate": 3.743530764807361e-05,
      "loss": 0.1171,
      "step": 5370
    },
    {
      "epoch": 1.4627514953779228,
      "grad_norm": 1.6328125,
      "learning_rate": 3.740655549166188e-05,
      "loss": 0.1529,
      "step": 5380
    },
    {
      "epoch": 1.4654703643284392,
      "grad_norm": 1.609375,
      "learning_rate": 3.7377803335250146e-05,
      "loss": 0.1229,
      "step": 5390
    },
    {
      "epoch": 1.468189233278956,
      "grad_norm": 2.09375,
      "learning_rate": 3.7349051178838414e-05,
      "loss": 0.1024,
      "step": 5400
    },
    {
      "epoch": 1.468189233278956,
      "eval_loss": 0.13067187368869781,
      "eval_runtime": 104.674,
      "eval_samples_per_second": 4.777,
      "eval_steps_per_second": 0.153,
      "step": 5400
    },
    {
      "epoch": 1.4709081022294725,
      "grad_norm": 1.1328125,
      "learning_rate": 3.732029902242669e-05,
      "loss": 0.1079,
      "step": 5410
    },
    {
      "epoch": 1.4736269711799892,
      "grad_norm": 1.4453125,
      "learning_rate": 3.7291546866014955e-05,
      "loss": 0.1208,
      "step": 5420
    },
    {
      "epoch": 1.4763458401305056,
      "grad_norm": 2.140625,
      "learning_rate": 3.726279470960322e-05,
      "loss": 0.1432,
      "step": 5430
    },
    {
      "epoch": 1.4790647090810223,
      "grad_norm": 1.125,
      "learning_rate": 3.723404255319149e-05,
      "loss": 0.1185,
      "step": 5440
    },
    {
      "epoch": 1.481783578031539,
      "grad_norm": 1.9296875,
      "learning_rate": 3.7205290396779764e-05,
      "loss": 0.1434,
      "step": 5450
    },
    {
      "epoch": 1.481783578031539,
      "eval_loss": 0.13780468702316284,
      "eval_runtime": 100.5858,
      "eval_samples_per_second": 4.971,
      "eval_steps_per_second": 0.159,
      "step": 5450
    },
    {
      "epoch": 1.4845024469820554,
      "grad_norm": 1.4375,
      "learning_rate": 3.717653824036803e-05,
      "loss": 0.1221,
      "step": 5460
    },
    {
      "epoch": 1.487221315932572,
      "grad_norm": 1.1875,
      "learning_rate": 3.71477860839563e-05,
      "loss": 0.1568,
      "step": 5470
    },
    {
      "epoch": 1.4899401848830887,
      "grad_norm": 2.09375,
      "learning_rate": 3.7119033927544566e-05,
      "loss": 0.1216,
      "step": 5480
    },
    {
      "epoch": 1.4926590538336053,
      "grad_norm": 2.484375,
      "learning_rate": 3.709028177113283e-05,
      "loss": 0.1203,
      "step": 5490
    },
    {
      "epoch": 1.4953779227841217,
      "grad_norm": 2.21875,
      "learning_rate": 3.706152961472111e-05,
      "loss": 0.1181,
      "step": 5500
    },
    {
      "epoch": 1.4953779227841217,
      "eval_loss": 0.12097655981779099,
      "eval_runtime": 100.3405,
      "eval_samples_per_second": 4.983,
      "eval_steps_per_second": 0.159,
      "step": 5500
    },
    {
      "epoch": 1.4980967917346384,
      "grad_norm": 2.390625,
      "learning_rate": 3.7032777458309374e-05,
      "loss": 0.1696,
      "step": 5510
    },
    {
      "epoch": 1.5008156606851548,
      "grad_norm": 1.9453125,
      "learning_rate": 3.700402530189764e-05,
      "loss": 0.1188,
      "step": 5520
    },
    {
      "epoch": 1.5035345296356715,
      "grad_norm": 2.09375,
      "learning_rate": 3.697527314548591e-05,
      "loss": 0.139,
      "step": 5530
    },
    {
      "epoch": 1.5062533985861881,
      "grad_norm": 2.796875,
      "learning_rate": 3.694652098907418e-05,
      "loss": 0.1529,
      "step": 5540
    },
    {
      "epoch": 1.5089722675367048,
      "grad_norm": 1.9375,
      "learning_rate": 3.691776883266245e-05,
      "loss": 0.1292,
      "step": 5550
    },
    {
      "epoch": 1.5089722675367048,
      "eval_loss": 0.13377343118190765,
      "eval_runtime": 98.4109,
      "eval_samples_per_second": 5.081,
      "eval_steps_per_second": 0.163,
      "step": 5550
    },
    {
      "epoch": 1.5116911364872214,
      "grad_norm": 2.578125,
      "learning_rate": 3.688901667625072e-05,
      "loss": 0.1318,
      "step": 5560
    },
    {
      "epoch": 1.5144100054377378,
      "grad_norm": 2.234375,
      "learning_rate": 3.6860264519838985e-05,
      "loss": 0.1256,
      "step": 5570
    },
    {
      "epoch": 1.5171288743882545,
      "grad_norm": 1.21875,
      "learning_rate": 3.683151236342726e-05,
      "loss": 0.1842,
      "step": 5580
    },
    {
      "epoch": 1.519847743338771,
      "grad_norm": 1.5390625,
      "learning_rate": 3.6802760207015527e-05,
      "loss": 0.1304,
      "step": 5590
    },
    {
      "epoch": 1.5225666122892876,
      "grad_norm": 1.3203125,
      "learning_rate": 3.67740080506038e-05,
      "loss": 0.119,
      "step": 5600
    },
    {
      "epoch": 1.5225666122892876,
      "eval_loss": 0.1298515647649765,
      "eval_runtime": 99.2411,
      "eval_samples_per_second": 5.038,
      "eval_steps_per_second": 0.161,
      "step": 5600
    },
    {
      "epoch": 1.5252854812398042,
      "grad_norm": 0.8359375,
      "learning_rate": 3.674525589419207e-05,
      "loss": 0.1188,
      "step": 5610
    },
    {
      "epoch": 1.5280043501903209,
      "grad_norm": 1.6953125,
      "learning_rate": 3.6716503737780335e-05,
      "loss": 0.113,
      "step": 5620
    },
    {
      "epoch": 1.5307232191408375,
      "grad_norm": 1.5078125,
      "learning_rate": 3.668775158136861e-05,
      "loss": 0.113,
      "step": 5630
    },
    {
      "epoch": 1.533442088091354,
      "grad_norm": 2.4375,
      "learning_rate": 3.665899942495688e-05,
      "loss": 0.1244,
      "step": 5640
    },
    {
      "epoch": 1.5361609570418706,
      "grad_norm": 1.0859375,
      "learning_rate": 3.6630247268545144e-05,
      "loss": 0.1259,
      "step": 5650
    },
    {
      "epoch": 1.5361609570418706,
      "eval_loss": 0.11897656321525574,
      "eval_runtime": 97.7344,
      "eval_samples_per_second": 5.116,
      "eval_steps_per_second": 0.164,
      "step": 5650
    },
    {
      "epoch": 1.538879825992387,
      "grad_norm": 2.5625,
      "learning_rate": 3.660149511213341e-05,
      "loss": 0.1317,
      "step": 5660
    },
    {
      "epoch": 1.5415986949429037,
      "grad_norm": 0.67578125,
      "learning_rate": 3.657274295572168e-05,
      "loss": 0.1123,
      "step": 5670
    },
    {
      "epoch": 1.5443175638934203,
      "grad_norm": 1.21875,
      "learning_rate": 3.654399079930995e-05,
      "loss": 0.1373,
      "step": 5680
    },
    {
      "epoch": 1.547036432843937,
      "grad_norm": 1.984375,
      "learning_rate": 3.651523864289822e-05,
      "loss": 0.1266,
      "step": 5690
    },
    {
      "epoch": 1.5497553017944536,
      "grad_norm": 1.84375,
      "learning_rate": 3.648648648648649e-05,
      "loss": 0.1136,
      "step": 5700
    },
    {
      "epoch": 1.5497553017944536,
      "eval_loss": 0.12717187404632568,
      "eval_runtime": 98.0066,
      "eval_samples_per_second": 5.102,
      "eval_steps_per_second": 0.163,
      "step": 5700
    },
    {
      "epoch": 1.55247417074497,
      "grad_norm": 1.5625,
      "learning_rate": 3.6457734330074755e-05,
      "loss": 0.1498,
      "step": 5710
    },
    {
      "epoch": 1.5551930396954867,
      "grad_norm": 1.0703125,
      "learning_rate": 3.642898217366303e-05,
      "loss": 0.1001,
      "step": 5720
    },
    {
      "epoch": 1.5579119086460032,
      "grad_norm": 0.82421875,
      "learning_rate": 3.6400230017251296e-05,
      "loss": 0.1115,
      "step": 5730
    },
    {
      "epoch": 1.5606307775965198,
      "grad_norm": 1.1875,
      "learning_rate": 3.637147786083956e-05,
      "loss": 0.1433,
      "step": 5740
    },
    {
      "epoch": 1.5633496465470365,
      "grad_norm": 1.5546875,
      "learning_rate": 3.634272570442783e-05,
      "loss": 0.15,
      "step": 5750
    },
    {
      "epoch": 1.5633496465470365,
      "eval_loss": 0.125914067029953,
      "eval_runtime": 97.8436,
      "eval_samples_per_second": 5.11,
      "eval_steps_per_second": 0.164,
      "step": 5750
    },
    {
      "epoch": 1.566068515497553,
      "grad_norm": 1.0078125,
      "learning_rate": 3.63139735480161e-05,
      "loss": 0.0991,
      "step": 5760
    },
    {
      "epoch": 1.5687873844480698,
      "grad_norm": 2.59375,
      "learning_rate": 3.628522139160437e-05,
      "loss": 0.1375,
      "step": 5770
    },
    {
      "epoch": 1.5715062533985862,
      "grad_norm": 1.828125,
      "learning_rate": 3.625646923519264e-05,
      "loss": 0.1203,
      "step": 5780
    },
    {
      "epoch": 1.5742251223491026,
      "grad_norm": 1.03125,
      "learning_rate": 3.622771707878091e-05,
      "loss": 0.1122,
      "step": 5790
    },
    {
      "epoch": 1.5769439912996193,
      "grad_norm": 1.4296875,
      "learning_rate": 3.6198964922369174e-05,
      "loss": 0.1139,
      "step": 5800
    },
    {
      "epoch": 1.5769439912996193,
      "eval_loss": 0.12578906118869781,
      "eval_runtime": 97.6714,
      "eval_samples_per_second": 5.119,
      "eval_steps_per_second": 0.164,
      "step": 5800
    },
    {
      "epoch": 1.579662860250136,
      "grad_norm": 2.015625,
      "learning_rate": 3.617021276595745e-05,
      "loss": 0.1118,
      "step": 5810
    },
    {
      "epoch": 1.5823817292006526,
      "grad_norm": 2.0625,
      "learning_rate": 3.6141460609545715e-05,
      "loss": 0.0989,
      "step": 5820
    },
    {
      "epoch": 1.5851005981511692,
      "grad_norm": 1.828125,
      "learning_rate": 3.611270845313399e-05,
      "loss": 0.1072,
      "step": 5830
    },
    {
      "epoch": 1.5878194671016856,
      "grad_norm": 2.03125,
      "learning_rate": 3.608395629672226e-05,
      "loss": 0.1153,
      "step": 5840
    },
    {
      "epoch": 1.5905383360522023,
      "grad_norm": 3.359375,
      "learning_rate": 3.6055204140310524e-05,
      "loss": 0.1163,
      "step": 5850
    },
    {
      "epoch": 1.5905383360522023,
      "eval_loss": 0.12254687398672104,
      "eval_runtime": 108.2444,
      "eval_samples_per_second": 4.619,
      "eval_steps_per_second": 0.148,
      "step": 5850
    },
    {
      "epoch": 1.5932572050027187,
      "grad_norm": 2.609375,
      "learning_rate": 3.60264519838988e-05,
      "loss": 0.1233,
      "step": 5860
    },
    {
      "epoch": 1.5959760739532354,
      "grad_norm": 1.421875,
      "learning_rate": 3.5997699827487066e-05,
      "loss": 0.1409,
      "step": 5870
    },
    {
      "epoch": 1.598694942903752,
      "grad_norm": 0.96484375,
      "learning_rate": 3.596894767107533e-05,
      "loss": 0.1058,
      "step": 5880
    },
    {
      "epoch": 1.6014138118542687,
      "grad_norm": 1.2578125,
      "learning_rate": 3.59401955146636e-05,
      "loss": 0.133,
      "step": 5890
    },
    {
      "epoch": 1.6041326808047853,
      "grad_norm": 2.453125,
      "learning_rate": 3.5911443358251874e-05,
      "loss": 0.0969,
      "step": 5900
    },
    {
      "epoch": 1.6041326808047853,
      "eval_loss": 0.1295781284570694,
      "eval_runtime": 100.3684,
      "eval_samples_per_second": 4.982,
      "eval_steps_per_second": 0.159,
      "step": 5900
    },
    {
      "epoch": 1.6068515497553018,
      "grad_norm": 1.453125,
      "learning_rate": 3.588269120184014e-05,
      "loss": 0.1042,
      "step": 5910
    },
    {
      "epoch": 1.6095704187058184,
      "grad_norm": 2.046875,
      "learning_rate": 3.585393904542841e-05,
      "loss": 0.13,
      "step": 5920
    },
    {
      "epoch": 1.6122892876563348,
      "grad_norm": 3.3125,
      "learning_rate": 3.5825186889016676e-05,
      "loss": 0.1262,
      "step": 5930
    },
    {
      "epoch": 1.6150081566068515,
      "grad_norm": 1.5,
      "learning_rate": 3.579643473260495e-05,
      "loss": 0.1256,
      "step": 5940
    },
    {
      "epoch": 1.6177270255573681,
      "grad_norm": 1.546875,
      "learning_rate": 3.576768257619322e-05,
      "loss": 0.1226,
      "step": 5950
    },
    {
      "epoch": 1.6177270255573681,
      "eval_loss": 0.12687499821186066,
      "eval_runtime": 99.5342,
      "eval_samples_per_second": 5.023,
      "eval_steps_per_second": 0.161,
      "step": 5950
    },
    {
      "epoch": 1.6204458945078848,
      "grad_norm": 1.1015625,
      "learning_rate": 3.5738930419781485e-05,
      "loss": 0.1407,
      "step": 5960
    },
    {
      "epoch": 1.6231647634584014,
      "grad_norm": 2.9375,
      "learning_rate": 3.571017826336975e-05,
      "loss": 0.1346,
      "step": 5970
    },
    {
      "epoch": 1.6258836324089179,
      "grad_norm": 1.9921875,
      "learning_rate": 3.568142610695802e-05,
      "loss": 0.1204,
      "step": 5980
    },
    {
      "epoch": 1.6286025013594345,
      "grad_norm": 0.921875,
      "learning_rate": 3.5652673950546294e-05,
      "loss": 0.0932,
      "step": 5990
    },
    {
      "epoch": 1.631321370309951,
      "grad_norm": 2.546875,
      "learning_rate": 3.562392179413456e-05,
      "loss": 0.1372,
      "step": 6000
    },
    {
      "epoch": 1.631321370309951,
      "eval_loss": 0.1318046897649765,
      "eval_runtime": 101.0327,
      "eval_samples_per_second": 4.949,
      "eval_steps_per_second": 0.158,
      "step": 6000
    },
    {
      "epoch": 1.6340402392604676,
      "grad_norm": 2.421875,
      "learning_rate": 3.559516963772283e-05,
      "loss": 0.1364,
      "step": 6010
    },
    {
      "epoch": 1.6367591082109842,
      "grad_norm": 2.375,
      "learning_rate": 3.5566417481311096e-05,
      "loss": 0.1451,
      "step": 6020
    },
    {
      "epoch": 1.639477977161501,
      "grad_norm": 0.95703125,
      "learning_rate": 3.553766532489937e-05,
      "loss": 0.1359,
      "step": 6030
    },
    {
      "epoch": 1.6421968461120175,
      "grad_norm": 2.171875,
      "learning_rate": 3.550891316848764e-05,
      "loss": 0.1481,
      "step": 6040
    },
    {
      "epoch": 1.644915715062534,
      "grad_norm": 0.8828125,
      "learning_rate": 3.5480161012075904e-05,
      "loss": 0.1452,
      "step": 6050
    },
    {
      "epoch": 1.644915715062534,
      "eval_loss": 0.1295468807220459,
      "eval_runtime": 97.7181,
      "eval_samples_per_second": 5.117,
      "eval_steps_per_second": 0.164,
      "step": 6050
    },
    {
      "epoch": 1.6476345840130504,
      "grad_norm": 1.2890625,
      "learning_rate": 3.545140885566418e-05,
      "loss": 0.1175,
      "step": 6060
    },
    {
      "epoch": 1.650353452963567,
      "grad_norm": 2.09375,
      "learning_rate": 3.5422656699252446e-05,
      "loss": 0.1392,
      "step": 6070
    },
    {
      "epoch": 1.6530723219140837,
      "grad_norm": 1.4609375,
      "learning_rate": 3.539390454284072e-05,
      "loss": 0.0969,
      "step": 6080
    },
    {
      "epoch": 1.6557911908646004,
      "grad_norm": 0.75390625,
      "learning_rate": 3.536515238642899e-05,
      "loss": 0.1059,
      "step": 6090
    },
    {
      "epoch": 1.658510059815117,
      "grad_norm": 3.265625,
      "learning_rate": 3.5336400230017254e-05,
      "loss": 0.1348,
      "step": 6100
    },
    {
      "epoch": 1.658510059815117,
      "eval_loss": 0.12873438000679016,
      "eval_runtime": 108.4768,
      "eval_samples_per_second": 4.609,
      "eval_steps_per_second": 0.147,
      "step": 6100
    },
    {
      "epoch": 1.6612289287656337,
      "grad_norm": 1.8125,
      "learning_rate": 3.530764807360552e-05,
      "loss": 0.1249,
      "step": 6110
    },
    {
      "epoch": 1.66394779771615,
      "grad_norm": 2.140625,
      "learning_rate": 3.5278895917193796e-05,
      "loss": 0.1438,
      "step": 6120
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 0.8671875,
      "learning_rate": 3.525014376078206e-05,
      "loss": 0.108,
      "step": 6130
    },
    {
      "epoch": 1.6693855356171832,
      "grad_norm": 1.9296875,
      "learning_rate": 3.522139160437033e-05,
      "loss": 0.1326,
      "step": 6140
    },
    {
      "epoch": 1.6721044045676998,
      "grad_norm": 1.5703125,
      "learning_rate": 3.51926394479586e-05,
      "loss": 0.1348,
      "step": 6150
    },
    {
      "epoch": 1.6721044045676998,
      "eval_loss": 0.11797656118869781,
      "eval_runtime": 100.6526,
      "eval_samples_per_second": 4.968,
      "eval_steps_per_second": 0.159,
      "step": 6150
    },
    {
      "epoch": 1.6748232735182165,
      "grad_norm": 1.578125,
      "learning_rate": 3.5163887291546865e-05,
      "loss": 0.1121,
      "step": 6160
    },
    {
      "epoch": 1.6775421424687331,
      "grad_norm": 0.91796875,
      "learning_rate": 3.513513513513514e-05,
      "loss": 0.099,
      "step": 6170
    },
    {
      "epoch": 1.6802610114192496,
      "grad_norm": 2.59375,
      "learning_rate": 3.5106382978723407e-05,
      "loss": 0.1364,
      "step": 6180
    },
    {
      "epoch": 1.6829798803697662,
      "grad_norm": 2.453125,
      "learning_rate": 3.5077630822311674e-05,
      "loss": 0.14,
      "step": 6190
    },
    {
      "epoch": 1.6856987493202826,
      "grad_norm": 3.375,
      "learning_rate": 3.504887866589994e-05,
      "loss": 0.1329,
      "step": 6200
    },
    {
      "epoch": 1.6856987493202826,
      "eval_loss": 0.1182265654206276,
      "eval_runtime": 97.1973,
      "eval_samples_per_second": 5.144,
      "eval_steps_per_second": 0.165,
      "step": 6200
    },
    {
      "epoch": 1.6884176182707993,
      "grad_norm": 1.796875,
      "learning_rate": 3.5020126509488215e-05,
      "loss": 0.1301,
      "step": 6210
    },
    {
      "epoch": 1.691136487221316,
      "grad_norm": 4.40625,
      "learning_rate": 3.499137435307648e-05,
      "loss": 0.1325,
      "step": 6220
    },
    {
      "epoch": 1.6938553561718326,
      "grad_norm": 2.328125,
      "learning_rate": 3.496262219666475e-05,
      "loss": 0.1513,
      "step": 6230
    },
    {
      "epoch": 1.6965742251223492,
      "grad_norm": 1.359375,
      "learning_rate": 3.493387004025302e-05,
      "loss": 0.0982,
      "step": 6240
    },
    {
      "epoch": 1.6992930940728657,
      "grad_norm": 2.28125,
      "learning_rate": 3.4905117883841284e-05,
      "loss": 0.106,
      "step": 6250
    },
    {
      "epoch": 1.6992930940728657,
      "eval_loss": 0.12956249713897705,
      "eval_runtime": 98.5028,
      "eval_samples_per_second": 5.076,
      "eval_steps_per_second": 0.162,
      "step": 6250
    },
    {
      "epoch": 1.7020119630233823,
      "grad_norm": 1.5234375,
      "learning_rate": 3.487636572742956e-05,
      "loss": 0.1148,
      "step": 6260
    },
    {
      "epoch": 1.7047308319738987,
      "grad_norm": 2.09375,
      "learning_rate": 3.4847613571017826e-05,
      "loss": 0.1483,
      "step": 6270
    },
    {
      "epoch": 1.7074497009244154,
      "grad_norm": 1.03125,
      "learning_rate": 3.481886141460609e-05,
      "loss": 0.0918,
      "step": 6280
    },
    {
      "epoch": 1.710168569874932,
      "grad_norm": 1.6875,
      "learning_rate": 3.479010925819437e-05,
      "loss": 0.109,
      "step": 6290
    },
    {
      "epoch": 1.7128874388254487,
      "grad_norm": 2.0625,
      "learning_rate": 3.4761357101782635e-05,
      "loss": 0.1548,
      "step": 6300
    },
    {
      "epoch": 1.7128874388254487,
      "eval_loss": 0.12278906255960464,
      "eval_runtime": 97.6879,
      "eval_samples_per_second": 5.118,
      "eval_steps_per_second": 0.164,
      "step": 6300
    },
    {
      "epoch": 1.7156063077759653,
      "grad_norm": 0.98046875,
      "learning_rate": 3.473260494537091e-05,
      "loss": 0.107,
      "step": 6310
    },
    {
      "epoch": 1.7183251767264818,
      "grad_norm": 1.1796875,
      "learning_rate": 3.4703852788959176e-05,
      "loss": 0.112,
      "step": 6320
    },
    {
      "epoch": 1.7210440456769984,
      "grad_norm": 1.203125,
      "learning_rate": 3.467510063254744e-05,
      "loss": 0.0912,
      "step": 6330
    },
    {
      "epoch": 1.7237629146275149,
      "grad_norm": 1.5234375,
      "learning_rate": 3.464634847613571e-05,
      "loss": 0.1278,
      "step": 6340
    },
    {
      "epoch": 1.7264817835780315,
      "grad_norm": 1.203125,
      "learning_rate": 3.4617596319723985e-05,
      "loss": 0.1349,
      "step": 6350
    },
    {
      "epoch": 1.7264817835780315,
      "eval_loss": 0.12380468845367432,
      "eval_runtime": 97.1213,
      "eval_samples_per_second": 5.148,
      "eval_steps_per_second": 0.165,
      "step": 6350
    },
    {
      "epoch": 1.7292006525285482,
      "grad_norm": 2.078125,
      "learning_rate": 3.458884416331225e-05,
      "loss": 0.1237,
      "step": 6360
    },
    {
      "epoch": 1.7319195214790648,
      "grad_norm": 1.390625,
      "learning_rate": 3.456009200690052e-05,
      "loss": 0.1116,
      "step": 6370
    },
    {
      "epoch": 1.7346383904295815,
      "grad_norm": 2.875,
      "learning_rate": 3.453133985048879e-05,
      "loss": 0.1326,
      "step": 6380
    },
    {
      "epoch": 1.7373572593800979,
      "grad_norm": 1.3359375,
      "learning_rate": 3.450258769407706e-05,
      "loss": 0.1194,
      "step": 6390
    },
    {
      "epoch": 1.7400761283306143,
      "grad_norm": 1.828125,
      "learning_rate": 3.447383553766533e-05,
      "loss": 0.1397,
      "step": 6400
    },
    {
      "epoch": 1.7400761283306143,
      "eval_loss": 0.12781250476837158,
      "eval_runtime": 104.8741,
      "eval_samples_per_second": 4.768,
      "eval_steps_per_second": 0.153,
      "step": 6400
    },
    {
      "epoch": 1.742794997281131,
      "grad_norm": 1.4140625,
      "learning_rate": 3.4445083381253595e-05,
      "loss": 0.1161,
      "step": 6410
    },
    {
      "epoch": 1.7455138662316476,
      "grad_norm": 1.859375,
      "learning_rate": 3.441633122484186e-05,
      "loss": 0.1376,
      "step": 6420
    },
    {
      "epoch": 1.7482327351821643,
      "grad_norm": 2.21875,
      "learning_rate": 3.438757906843013e-05,
      "loss": 0.147,
      "step": 6430
    },
    {
      "epoch": 1.750951604132681,
      "grad_norm": 1.7890625,
      "learning_rate": 3.4358826912018404e-05,
      "loss": 0.1225,
      "step": 6440
    },
    {
      "epoch": 1.7536704730831973,
      "grad_norm": 1.4296875,
      "learning_rate": 3.433007475560667e-05,
      "loss": 0.1552,
      "step": 6450
    },
    {
      "epoch": 1.7536704730831973,
      "eval_loss": 0.12250000238418579,
      "eval_runtime": 98.6077,
      "eval_samples_per_second": 5.071,
      "eval_steps_per_second": 0.162,
      "step": 6450
    },
    {
      "epoch": 1.756389342033714,
      "grad_norm": 1.421875,
      "learning_rate": 3.430132259919494e-05,
      "loss": 0.1395,
      "step": 6460
    },
    {
      "epoch": 1.7591082109842304,
      "grad_norm": 2.140625,
      "learning_rate": 3.4272570442783206e-05,
      "loss": 0.1168,
      "step": 6470
    },
    {
      "epoch": 1.761827079934747,
      "grad_norm": 1.5,
      "learning_rate": 3.424381828637148e-05,
      "loss": 0.1096,
      "step": 6480
    },
    {
      "epoch": 1.7645459488852637,
      "grad_norm": 2.328125,
      "learning_rate": 3.421506612995975e-05,
      "loss": 0.1371,
      "step": 6490
    },
    {
      "epoch": 1.7672648178357804,
      "grad_norm": 4.34375,
      "learning_rate": 3.4186313973548015e-05,
      "loss": 0.1278,
      "step": 6500
    },
    {
      "epoch": 1.7672648178357804,
      "eval_loss": 0.12501563131809235,
      "eval_runtime": 99.6979,
      "eval_samples_per_second": 5.015,
      "eval_steps_per_second": 0.16,
      "step": 6500
    },
    {
      "epoch": 1.769983686786297,
      "grad_norm": 2.1875,
      "learning_rate": 3.415756181713628e-05,
      "loss": 0.1226,
      "step": 6510
    },
    {
      "epoch": 1.7727025557368135,
      "grad_norm": 1.59375,
      "learning_rate": 3.4128809660724556e-05,
      "loss": 0.099,
      "step": 6520
    },
    {
      "epoch": 1.77542142468733,
      "grad_norm": 1.171875,
      "learning_rate": 3.4100057504312824e-05,
      "loss": 0.1235,
      "step": 6530
    },
    {
      "epoch": 1.7781402936378465,
      "grad_norm": 1.296875,
      "learning_rate": 3.40713053479011e-05,
      "loss": 0.1271,
      "step": 6540
    },
    {
      "epoch": 1.7808591625883632,
      "grad_norm": 1.703125,
      "learning_rate": 3.4042553191489365e-05,
      "loss": 0.1276,
      "step": 6550
    },
    {
      "epoch": 1.7808591625883632,
      "eval_loss": 0.12169530987739563,
      "eval_runtime": 99.0483,
      "eval_samples_per_second": 5.048,
      "eval_steps_per_second": 0.162,
      "step": 6550
    },
    {
      "epoch": 1.7835780315388798,
      "grad_norm": 1.359375,
      "learning_rate": 3.401380103507763e-05,
      "loss": 0.113,
      "step": 6560
    },
    {
      "epoch": 1.7862969004893965,
      "grad_norm": 1.3125,
      "learning_rate": 3.3985048878665906e-05,
      "loss": 0.1088,
      "step": 6570
    },
    {
      "epoch": 1.7890157694399131,
      "grad_norm": 1.3359375,
      "learning_rate": 3.3956296722254174e-05,
      "loss": 0.1205,
      "step": 6580
    },
    {
      "epoch": 1.7917346383904296,
      "grad_norm": 1.578125,
      "learning_rate": 3.392754456584244e-05,
      "loss": 0.126,
      "step": 6590
    },
    {
      "epoch": 1.7944535073409462,
      "grad_norm": 1.375,
      "learning_rate": 3.389879240943071e-05,
      "loss": 0.1138,
      "step": 6600
    },
    {
      "epoch": 1.7944535073409462,
      "eval_loss": 0.1284531205892563,
      "eval_runtime": 97.7985,
      "eval_samples_per_second": 5.113,
      "eval_steps_per_second": 0.164,
      "step": 6600
    },
    {
      "epoch": 1.7971723762914626,
      "grad_norm": 2.71875,
      "learning_rate": 3.3870040253018976e-05,
      "loss": 0.1244,
      "step": 6610
    },
    {
      "epoch": 1.7998912452419793,
      "grad_norm": 1.703125,
      "learning_rate": 3.384128809660725e-05,
      "loss": 0.1153,
      "step": 6620
    },
    {
      "epoch": 1.802610114192496,
      "grad_norm": 2.1875,
      "learning_rate": 3.381253594019552e-05,
      "loss": 0.1324,
      "step": 6630
    },
    {
      "epoch": 1.8053289831430126,
      "grad_norm": 1.375,
      "learning_rate": 3.3783783783783784e-05,
      "loss": 0.1064,
      "step": 6640
    },
    {
      "epoch": 1.8080478520935293,
      "grad_norm": 3.140625,
      "learning_rate": 3.375503162737205e-05,
      "loss": 0.1084,
      "step": 6650
    },
    {
      "epoch": 1.8080478520935293,
      "eval_loss": 0.12862500548362732,
      "eval_runtime": 97.0131,
      "eval_samples_per_second": 5.154,
      "eval_steps_per_second": 0.165,
      "step": 6650
    },
    {
      "epoch": 1.8107667210440457,
      "grad_norm": 1.3515625,
      "learning_rate": 3.3726279470960326e-05,
      "loss": 0.0922,
      "step": 6660
    },
    {
      "epoch": 1.813485589994562,
      "grad_norm": 1.8828125,
      "learning_rate": 3.369752731454859e-05,
      "loss": 0.1712,
      "step": 6670
    },
    {
      "epoch": 1.8162044589450788,
      "grad_norm": 1.6796875,
      "learning_rate": 3.366877515813686e-05,
      "loss": 0.144,
      "step": 6680
    },
    {
      "epoch": 1.8189233278955954,
      "grad_norm": 2.09375,
      "learning_rate": 3.364002300172513e-05,
      "loss": 0.119,
      "step": 6690
    },
    {
      "epoch": 1.821642196846112,
      "grad_norm": 2.359375,
      "learning_rate": 3.36112708453134e-05,
      "loss": 0.1154,
      "step": 6700
    },
    {
      "epoch": 1.821642196846112,
      "eval_loss": 0.12128905951976776,
      "eval_runtime": 97.8772,
      "eval_samples_per_second": 5.108,
      "eval_steps_per_second": 0.163,
      "step": 6700
    },
    {
      "epoch": 1.8243610657966287,
      "grad_norm": 0.9375,
      "learning_rate": 3.358251868890167e-05,
      "loss": 0.1212,
      "step": 6710
    },
    {
      "epoch": 1.8270799347471451,
      "grad_norm": 1.578125,
      "learning_rate": 3.3553766532489936e-05,
      "loss": 0.1064,
      "step": 6720
    },
    {
      "epoch": 1.8297988036976618,
      "grad_norm": 2.078125,
      "learning_rate": 3.3525014376078204e-05,
      "loss": 0.1318,
      "step": 6730
    },
    {
      "epoch": 1.8325176726481782,
      "grad_norm": 1.7890625,
      "learning_rate": 3.349626221966647e-05,
      "loss": 0.0988,
      "step": 6740
    },
    {
      "epoch": 1.8352365415986949,
      "grad_norm": 1.3203125,
      "learning_rate": 3.3467510063254745e-05,
      "loss": 0.1077,
      "step": 6750
    },
    {
      "epoch": 1.8352365415986949,
      "eval_loss": 0.12403906136751175,
      "eval_runtime": 96.0624,
      "eval_samples_per_second": 5.205,
      "eval_steps_per_second": 0.167,
      "step": 6750
    },
    {
      "epoch": 1.8379554105492115,
      "grad_norm": 1.3671875,
      "learning_rate": 3.343875790684301e-05,
      "loss": 0.1293,
      "step": 6760
    },
    {
      "epoch": 1.8406742794997282,
      "grad_norm": 2.65625,
      "learning_rate": 3.3410005750431286e-05,
      "loss": 0.1163,
      "step": 6770
    },
    {
      "epoch": 1.8433931484502448,
      "grad_norm": 2.125,
      "learning_rate": 3.3381253594019554e-05,
      "loss": 0.16,
      "step": 6780
    },
    {
      "epoch": 1.8461120174007613,
      "grad_norm": 2.0625,
      "learning_rate": 3.335250143760783e-05,
      "loss": 0.1355,
      "step": 6790
    },
    {
      "epoch": 1.848830886351278,
      "grad_norm": 0.56640625,
      "learning_rate": 3.3323749281196095e-05,
      "loss": 0.0953,
      "step": 6800
    },
    {
      "epoch": 1.848830886351278,
      "eval_loss": 0.11652343720197678,
      "eval_runtime": 99.3208,
      "eval_samples_per_second": 5.034,
      "eval_steps_per_second": 0.161,
      "step": 6800
    },
    {
      "epoch": 1.8515497553017943,
      "grad_norm": 1.9296875,
      "learning_rate": 3.329499712478436e-05,
      "loss": 0.104,
      "step": 6810
    },
    {
      "epoch": 1.854268624252311,
      "grad_norm": 1.5,
      "learning_rate": 3.326624496837263e-05,
      "loss": 0.1168,
      "step": 6820
    },
    {
      "epoch": 1.8569874932028276,
      "grad_norm": 2.015625,
      "learning_rate": 3.32374928119609e-05,
      "loss": 0.1197,
      "step": 6830
    },
    {
      "epoch": 1.8597063621533443,
      "grad_norm": 0.73046875,
      "learning_rate": 3.320874065554917e-05,
      "loss": 0.1182,
      "step": 6840
    },
    {
      "epoch": 1.862425231103861,
      "grad_norm": 1.171875,
      "learning_rate": 3.317998849913744e-05,
      "loss": 0.1252,
      "step": 6850
    },
    {
      "epoch": 1.862425231103861,
      "eval_loss": 0.11440625041723251,
      "eval_runtime": 99.2412,
      "eval_samples_per_second": 5.038,
      "eval_steps_per_second": 0.161,
      "step": 6850
    },
    {
      "epoch": 1.8651441000543774,
      "grad_norm": 1.078125,
      "learning_rate": 3.3151236342725706e-05,
      "loss": 0.128,
      "step": 6860
    },
    {
      "epoch": 1.867862969004894,
      "grad_norm": 2.25,
      "learning_rate": 3.312248418631397e-05,
      "loss": 0.1207,
      "step": 6870
    },
    {
      "epoch": 1.8705818379554104,
      "grad_norm": 1.6328125,
      "learning_rate": 3.309373202990225e-05,
      "loss": 0.1418,
      "step": 6880
    },
    {
      "epoch": 1.873300706905927,
      "grad_norm": 5.5,
      "learning_rate": 3.3064979873490515e-05,
      "loss": 0.1407,
      "step": 6890
    },
    {
      "epoch": 1.8760195758564437,
      "grad_norm": 1.484375,
      "learning_rate": 3.303622771707878e-05,
      "loss": 0.1425,
      "step": 6900
    },
    {
      "epoch": 1.8760195758564437,
      "eval_loss": 0.1152656227350235,
      "eval_runtime": 180.5883,
      "eval_samples_per_second": 2.769,
      "eval_steps_per_second": 0.089,
      "step": 6900
    },
    {
      "epoch": 1.8787384448069604,
      "grad_norm": 1.3984375,
      "learning_rate": 3.300747556066705e-05,
      "loss": 0.111,
      "step": 6910
    },
    {
      "epoch": 1.881457313757477,
      "grad_norm": 1.2578125,
      "learning_rate": 3.2978723404255317e-05,
      "loss": 0.1025,
      "step": 6920
    },
    {
      "epoch": 1.8841761827079935,
      "grad_norm": 1.3828125,
      "learning_rate": 3.294997124784359e-05,
      "loss": 0.1244,
      "step": 6930
    },
    {
      "epoch": 1.88689505165851,
      "grad_norm": 2.546875,
      "learning_rate": 3.292121909143186e-05,
      "loss": 0.1111,
      "step": 6940
    },
    {
      "epoch": 1.8896139206090266,
      "grad_norm": 1.7890625,
      "learning_rate": 3.2892466935020125e-05,
      "loss": 0.1289,
      "step": 6950
    },
    {
      "epoch": 1.8896139206090266,
      "eval_loss": 0.12961718440055847,
      "eval_runtime": 104.6145,
      "eval_samples_per_second": 4.779,
      "eval_steps_per_second": 0.153,
      "step": 6950
    },
    {
      "epoch": 1.8923327895595432,
      "grad_norm": 1.8359375,
      "learning_rate": 3.286371477860839e-05,
      "loss": 0.1045,
      "step": 6960
    },
    {
      "epoch": 1.8950516585100599,
      "grad_norm": 2.359375,
      "learning_rate": 3.283496262219667e-05,
      "loss": 0.1489,
      "step": 6970
    },
    {
      "epoch": 1.8977705274605765,
      "grad_norm": 2.671875,
      "learning_rate": 3.2806210465784934e-05,
      "loss": 0.1065,
      "step": 6980
    },
    {
      "epoch": 1.900489396411093,
      "grad_norm": 2.09375,
      "learning_rate": 3.27774583093732e-05,
      "loss": 0.1258,
      "step": 6990
    },
    {
      "epoch": 1.9032082653616096,
      "grad_norm": 2.078125,
      "learning_rate": 3.2748706152961475e-05,
      "loss": 0.1313,
      "step": 7000
    },
    {
      "epoch": 1.9032082653616096,
      "eval_loss": 0.1376015692949295,
      "eval_runtime": 100.1736,
      "eval_samples_per_second": 4.991,
      "eval_steps_per_second": 0.16,
      "step": 7000
    },
    {
      "epoch": 1.905927134312126,
      "grad_norm": 3.109375,
      "learning_rate": 3.271995399654974e-05,
      "loss": 0.1344,
      "step": 7010
    },
    {
      "epoch": 1.9086460032626427,
      "grad_norm": 2.140625,
      "learning_rate": 3.269120184013802e-05,
      "loss": 0.103,
      "step": 7020
    },
    {
      "epoch": 1.9113648722131593,
      "grad_norm": 1.984375,
      "learning_rate": 3.2662449683726284e-05,
      "loss": 0.1042,
      "step": 7030
    },
    {
      "epoch": 1.914083741163676,
      "grad_norm": 2.15625,
      "learning_rate": 3.263369752731455e-05,
      "loss": 0.092,
      "step": 7040
    },
    {
      "epoch": 1.9168026101141926,
      "grad_norm": 1.546875,
      "learning_rate": 3.260494537090282e-05,
      "loss": 0.1268,
      "step": 7050
    },
    {
      "epoch": 1.9168026101141926,
      "eval_loss": 0.1217968761920929,
      "eval_runtime": 99.6632,
      "eval_samples_per_second": 5.017,
      "eval_steps_per_second": 0.161,
      "step": 7050
    },
    {
      "epoch": 1.919521479064709,
      "grad_norm": 1.515625,
      "learning_rate": 3.257619321449109e-05,
      "loss": 0.1221,
      "step": 7060
    },
    {
      "epoch": 1.9222403480152257,
      "grad_norm": 1.578125,
      "learning_rate": 3.254744105807936e-05,
      "loss": 0.1076,
      "step": 7070
    },
    {
      "epoch": 1.9249592169657421,
      "grad_norm": 1.78125,
      "learning_rate": 3.251868890166763e-05,
      "loss": 0.1382,
      "step": 7080
    },
    {
      "epoch": 1.9276780859162588,
      "grad_norm": 0.984375,
      "learning_rate": 3.2489936745255895e-05,
      "loss": 0.1294,
      "step": 7090
    },
    {
      "epoch": 1.9303969548667754,
      "grad_norm": 1.8359375,
      "learning_rate": 3.246118458884416e-05,
      "loss": 0.1193,
      "step": 7100
    },
    {
      "epoch": 1.9303969548667754,
      "eval_loss": 0.12407030910253525,
      "eval_runtime": 101.0603,
      "eval_samples_per_second": 4.948,
      "eval_steps_per_second": 0.158,
      "step": 7100
    },
    {
      "epoch": 1.933115823817292,
      "grad_norm": 1.7890625,
      "learning_rate": 3.2432432432432436e-05,
      "loss": 0.1346,
      "step": 7110
    },
    {
      "epoch": 1.9358346927678087,
      "grad_norm": 1.6015625,
      "learning_rate": 3.2403680276020703e-05,
      "loss": 0.109,
      "step": 7120
    },
    {
      "epoch": 1.9385535617183252,
      "grad_norm": 3.21875,
      "learning_rate": 3.237492811960897e-05,
      "loss": 0.1327,
      "step": 7130
    },
    {
      "epoch": 1.9412724306688418,
      "grad_norm": 0.86328125,
      "learning_rate": 3.234617596319724e-05,
      "loss": 0.1087,
      "step": 7140
    },
    {
      "epoch": 1.9439912996193582,
      "grad_norm": 1.84375,
      "learning_rate": 3.231742380678551e-05,
      "loss": 0.1226,
      "step": 7150
    },
    {
      "epoch": 1.9439912996193582,
      "eval_loss": 0.12222656607627869,
      "eval_runtime": 99.2834,
      "eval_samples_per_second": 5.036,
      "eval_steps_per_second": 0.161,
      "step": 7150
    },
    {
      "epoch": 1.9467101685698749,
      "grad_norm": 1.21875,
      "learning_rate": 3.228867165037378e-05,
      "loss": 0.1244,
      "step": 7160
    },
    {
      "epoch": 1.9494290375203915,
      "grad_norm": 2.703125,
      "learning_rate": 3.225991949396205e-05,
      "loss": 0.1417,
      "step": 7170
    },
    {
      "epoch": 1.9521479064709082,
      "grad_norm": 1.5078125,
      "learning_rate": 3.2231167337550314e-05,
      "loss": 0.0969,
      "step": 7180
    },
    {
      "epoch": 1.9548667754214248,
      "grad_norm": 0.828125,
      "learning_rate": 3.220241518113858e-05,
      "loss": 0.1052,
      "step": 7190
    },
    {
      "epoch": 1.9575856443719413,
      "grad_norm": 2.296875,
      "learning_rate": 3.2173663024726856e-05,
      "loss": 0.144,
      "step": 7200
    },
    {
      "epoch": 1.9575856443719413,
      "eval_loss": 0.12021874636411667,
      "eval_runtime": 96.8549,
      "eval_samples_per_second": 5.162,
      "eval_steps_per_second": 0.165,
      "step": 7200
    },
    {
      "epoch": 1.9603045133224577,
      "grad_norm": 0.953125,
      "learning_rate": 3.214491086831512e-05,
      "loss": 0.1372,
      "step": 7210
    },
    {
      "epoch": 1.9630233822729743,
      "grad_norm": 1.15625,
      "learning_rate": 3.211615871190339e-05,
      "loss": 0.1169,
      "step": 7220
    },
    {
      "epoch": 1.965742251223491,
      "grad_norm": 2.140625,
      "learning_rate": 3.2087406555491664e-05,
      "loss": 0.1687,
      "step": 7230
    },
    {
      "epoch": 1.9684611201740077,
      "grad_norm": 1.0546875,
      "learning_rate": 3.205865439907993e-05,
      "loss": 0.152,
      "step": 7240
    },
    {
      "epoch": 1.9711799891245243,
      "grad_norm": 1.921875,
      "learning_rate": 3.2029902242668206e-05,
      "loss": 0.0976,
      "step": 7250
    },
    {
      "epoch": 1.9711799891245243,
      "eval_loss": 0.11679687350988388,
      "eval_runtime": 98.0673,
      "eval_samples_per_second": 5.099,
      "eval_steps_per_second": 0.163,
      "step": 7250
    },
    {
      "epoch": 1.9738988580750407,
      "grad_norm": 2.78125,
      "learning_rate": 3.200115008625647e-05,
      "loss": 0.1298,
      "step": 7260
    },
    {
      "epoch": 1.9766177270255574,
      "grad_norm": 1.5390625,
      "learning_rate": 3.197239792984474e-05,
      "loss": 0.1563,
      "step": 7270
    },
    {
      "epoch": 1.9793365959760738,
      "grad_norm": 0.92578125,
      "learning_rate": 3.194364577343301e-05,
      "loss": 0.1415,
      "step": 7280
    },
    {
      "epoch": 1.9820554649265905,
      "grad_norm": 1.1171875,
      "learning_rate": 3.191489361702128e-05,
      "loss": 0.0943,
      "step": 7290
    },
    {
      "epoch": 1.984774333877107,
      "grad_norm": 1.75,
      "learning_rate": 3.188614146060955e-05,
      "loss": 0.1375,
      "step": 7300
    },
    {
      "epoch": 1.984774333877107,
      "eval_loss": 0.10882812738418579,
      "eval_runtime": 96.8984,
      "eval_samples_per_second": 5.16,
      "eval_steps_per_second": 0.165,
      "step": 7300
    },
    {
      "epoch": 1.9874932028276238,
      "grad_norm": 1.2578125,
      "learning_rate": 3.1857389304197816e-05,
      "loss": 0.1103,
      "step": 7310
    },
    {
      "epoch": 1.9902120717781404,
      "grad_norm": 1.859375,
      "learning_rate": 3.1828637147786084e-05,
      "loss": 0.1004,
      "step": 7320
    },
    {
      "epoch": 1.9929309407286568,
      "grad_norm": 1.6171875,
      "learning_rate": 3.179988499137436e-05,
      "loss": 0.1163,
      "step": 7330
    },
    {
      "epoch": 1.9956498096791735,
      "grad_norm": 1.40625,
      "learning_rate": 3.1771132834962625e-05,
      "loss": 0.1216,
      "step": 7340
    },
    {
      "epoch": 1.99836867862969,
      "grad_norm": 1.34375,
      "learning_rate": 3.174238067855089e-05,
      "loss": 0.1105,
      "step": 7350
    },
    {
      "epoch": 1.99836867862969,
      "eval_loss": 0.12134374678134918,
      "eval_runtime": 108.8497,
      "eval_samples_per_second": 4.593,
      "eval_steps_per_second": 0.147,
      "step": 7350
    },
    {
      "epoch": 2.0010875475802066,
      "grad_norm": 1.4921875,
      "learning_rate": 3.171362852213916e-05,
      "loss": 0.1034,
      "step": 7360
    },
    {
      "epoch": 2.0038064165307232,
      "grad_norm": 2.140625,
      "learning_rate": 3.168487636572743e-05,
      "loss": 0.104,
      "step": 7370
    },
    {
      "epoch": 2.00652528548124,
      "grad_norm": 1.34375,
      "learning_rate": 3.16561242093157e-05,
      "loss": 0.0951,
      "step": 7380
    },
    {
      "epoch": 2.0092441544317565,
      "grad_norm": 1.1015625,
      "learning_rate": 3.162737205290397e-05,
      "loss": 0.1135,
      "step": 7390
    },
    {
      "epoch": 2.011963023382273,
      "grad_norm": 0.94921875,
      "learning_rate": 3.1598619896492236e-05,
      "loss": 0.0939,
      "step": 7400
    },
    {
      "epoch": 2.011963023382273,
      "eval_loss": 0.1291171908378601,
      "eval_runtime": 99.0925,
      "eval_samples_per_second": 5.046,
      "eval_steps_per_second": 0.161,
      "step": 7400
    },
    {
      "epoch": 2.0146818923327894,
      "grad_norm": 1.1640625,
      "learning_rate": 3.15698677400805e-05,
      "loss": 0.0998,
      "step": 7410
    },
    {
      "epoch": 2.017400761283306,
      "grad_norm": 1.09375,
      "learning_rate": 3.154111558366878e-05,
      "loss": 0.0966,
      "step": 7420
    },
    {
      "epoch": 2.0201196302338227,
      "grad_norm": 1.7421875,
      "learning_rate": 3.1512363427257044e-05,
      "loss": 0.1148,
      "step": 7430
    },
    {
      "epoch": 2.0228384991843393,
      "grad_norm": 0.8046875,
      "learning_rate": 3.148361127084531e-05,
      "loss": 0.0915,
      "step": 7440
    },
    {
      "epoch": 2.025557368134856,
      "grad_norm": 0.98828125,
      "learning_rate": 3.145485911443358e-05,
      "loss": 0.0871,
      "step": 7450
    },
    {
      "epoch": 2.025557368134856,
      "eval_loss": 0.12837499380111694,
      "eval_runtime": 99.5011,
      "eval_samples_per_second": 5.025,
      "eval_steps_per_second": 0.161,
      "step": 7450
    },
    {
      "epoch": 2.0282762370853726,
      "grad_norm": 1.1171875,
      "learning_rate": 3.142610695802185e-05,
      "loss": 0.0913,
      "step": 7460
    },
    {
      "epoch": 2.0309951060358893,
      "grad_norm": 3.703125,
      "learning_rate": 3.139735480161012e-05,
      "loss": 0.1101,
      "step": 7470
    },
    {
      "epoch": 2.0337139749864055,
      "grad_norm": 3.5,
      "learning_rate": 3.1368602645198395e-05,
      "loss": 0.1084,
      "step": 7480
    },
    {
      "epoch": 2.036432843936922,
      "grad_norm": 2.75,
      "learning_rate": 3.133985048878666e-05,
      "loss": 0.1073,
      "step": 7490
    },
    {
      "epoch": 2.039151712887439,
      "grad_norm": 1.3203125,
      "learning_rate": 3.131109833237493e-05,
      "loss": 0.0779,
      "step": 7500
    },
    {
      "epoch": 2.039151712887439,
      "eval_loss": 0.10894531011581421,
      "eval_runtime": 97.5213,
      "eval_samples_per_second": 5.127,
      "eval_steps_per_second": 0.164,
      "step": 7500
    },
    {
      "epoch": 2.0418705818379554,
      "grad_norm": 1.6796875,
      "learning_rate": 3.12823461759632e-05,
      "loss": 0.1153,
      "step": 7510
    },
    {
      "epoch": 2.044589450788472,
      "grad_norm": 2.515625,
      "learning_rate": 3.125359401955147e-05,
      "loss": 0.0959,
      "step": 7520
    },
    {
      "epoch": 2.0473083197389887,
      "grad_norm": 1.8984375,
      "learning_rate": 3.122484186313974e-05,
      "loss": 0.1129,
      "step": 7530
    },
    {
      "epoch": 2.050027188689505,
      "grad_norm": 1.328125,
      "learning_rate": 3.1196089706728005e-05,
      "loss": 0.0901,
      "step": 7540
    },
    {
      "epoch": 2.0527460576400216,
      "grad_norm": 1.5703125,
      "learning_rate": 3.116733755031628e-05,
      "loss": 0.1283,
      "step": 7550
    },
    {
      "epoch": 2.0527460576400216,
      "eval_loss": 0.11101562529802322,
      "eval_runtime": 97.9469,
      "eval_samples_per_second": 5.105,
      "eval_steps_per_second": 0.163,
      "step": 7550
    },
    {
      "epoch": 2.0554649265905383,
      "grad_norm": 1.4375,
      "learning_rate": 3.113858539390455e-05,
      "loss": 0.1172,
      "step": 7560
    },
    {
      "epoch": 2.058183795541055,
      "grad_norm": 1.328125,
      "learning_rate": 3.1109833237492814e-05,
      "loss": 0.1059,
      "step": 7570
    },
    {
      "epoch": 2.0609026644915716,
      "grad_norm": 0.94921875,
      "learning_rate": 3.108108108108108e-05,
      "loss": 0.1039,
      "step": 7580
    },
    {
      "epoch": 2.063621533442088,
      "grad_norm": 0.8515625,
      "learning_rate": 3.105232892466935e-05,
      "loss": 0.0939,
      "step": 7590
    },
    {
      "epoch": 2.066340402392605,
      "grad_norm": 2.265625,
      "learning_rate": 3.102357676825762e-05,
      "loss": 0.1145,
      "step": 7600
    },
    {
      "epoch": 2.066340402392605,
      "eval_loss": 0.11525781452655792,
      "eval_runtime": 119.0563,
      "eval_samples_per_second": 4.2,
      "eval_steps_per_second": 0.134,
      "step": 7600
    },
    {
      "epoch": 2.069059271343121,
      "grad_norm": 1.03125,
      "learning_rate": 3.099482461184589e-05,
      "loss": 0.114,
      "step": 7610
    },
    {
      "epoch": 2.0717781402936377,
      "grad_norm": 1.4296875,
      "learning_rate": 3.096607245543416e-05,
      "loss": 0.0984,
      "step": 7620
    },
    {
      "epoch": 2.0744970092441544,
      "grad_norm": 1.15625,
      "learning_rate": 3.0937320299022425e-05,
      "loss": 0.1041,
      "step": 7630
    },
    {
      "epoch": 2.077215878194671,
      "grad_norm": 1.4453125,
      "learning_rate": 3.09085681426107e-05,
      "loss": 0.0825,
      "step": 7640
    },
    {
      "epoch": 2.0799347471451877,
      "grad_norm": 1.6953125,
      "learning_rate": 3.0879815986198966e-05,
      "loss": 0.1334,
      "step": 7650
    },
    {
      "epoch": 2.0799347471451877,
      "eval_loss": 0.12605468928813934,
      "eval_runtime": 100.6244,
      "eval_samples_per_second": 4.969,
      "eval_steps_per_second": 0.159,
      "step": 7650
    },
    {
      "epoch": 2.0826536160957043,
      "grad_norm": 1.2578125,
      "learning_rate": 3.085106382978723e-05,
      "loss": 0.0971,
      "step": 7660
    },
    {
      "epoch": 2.085372485046221,
      "grad_norm": 3.21875,
      "learning_rate": 3.08223116733755e-05,
      "loss": 0.1141,
      "step": 7670
    },
    {
      "epoch": 2.088091353996737,
      "grad_norm": 1.3125,
      "learning_rate": 3.0793559516963775e-05,
      "loss": 0.117,
      "step": 7680
    },
    {
      "epoch": 2.090810222947254,
      "grad_norm": 1.09375,
      "learning_rate": 3.076480736055204e-05,
      "loss": 0.1036,
      "step": 7690
    },
    {
      "epoch": 2.0935290918977705,
      "grad_norm": 1.7265625,
      "learning_rate": 3.073605520414031e-05,
      "loss": 0.1079,
      "step": 7700
    },
    {
      "epoch": 2.0935290918977705,
      "eval_loss": 0.12360156327486038,
      "eval_runtime": 96.6715,
      "eval_samples_per_second": 5.172,
      "eval_steps_per_second": 0.166,
      "step": 7700
    },
    {
      "epoch": 2.096247960848287,
      "grad_norm": 2.359375,
      "learning_rate": 3.0707303047728583e-05,
      "loss": 0.0999,
      "step": 7710
    },
    {
      "epoch": 2.098966829798804,
      "grad_norm": 1.2265625,
      "learning_rate": 3.067855089131685e-05,
      "loss": 0.1135,
      "step": 7720
    },
    {
      "epoch": 2.1016856987493204,
      "grad_norm": 1.453125,
      "learning_rate": 3.0649798734905125e-05,
      "loss": 0.1035,
      "step": 7730
    },
    {
      "epoch": 2.104404567699837,
      "grad_norm": 1.5078125,
      "learning_rate": 3.062104657849339e-05,
      "loss": 0.1074,
      "step": 7740
    },
    {
      "epoch": 2.1071234366503533,
      "grad_norm": 1.625,
      "learning_rate": 3.059229442208166e-05,
      "loss": 0.1033,
      "step": 7750
    },
    {
      "epoch": 2.1071234366503533,
      "eval_loss": 0.1220703125,
      "eval_runtime": 95.3496,
      "eval_samples_per_second": 5.244,
      "eval_steps_per_second": 0.168,
      "step": 7750
    },
    {
      "epoch": 2.10984230560087,
      "grad_norm": 2.4375,
      "learning_rate": 3.056354226566993e-05,
      "loss": 0.121,
      "step": 7760
    },
    {
      "epoch": 2.1125611745513866,
      "grad_norm": 0.75390625,
      "learning_rate": 3.0534790109258194e-05,
      "loss": 0.0822,
      "step": 7770
    },
    {
      "epoch": 2.1152800435019032,
      "grad_norm": 1.28125,
      "learning_rate": 3.0506037952846468e-05,
      "loss": 0.12,
      "step": 7780
    },
    {
      "epoch": 2.11799891245242,
      "grad_norm": 7.4375,
      "learning_rate": 3.0477285796434736e-05,
      "loss": 0.1166,
      "step": 7790
    },
    {
      "epoch": 2.1207177814029365,
      "grad_norm": 1.6875,
      "learning_rate": 3.0448533640023003e-05,
      "loss": 0.0794,
      "step": 7800
    },
    {
      "epoch": 2.1207177814029365,
      "eval_loss": 0.1296093761920929,
      "eval_runtime": 96.3835,
      "eval_samples_per_second": 5.188,
      "eval_steps_per_second": 0.166,
      "step": 7800
    },
    {
      "epoch": 2.1234366503534527,
      "grad_norm": 1.390625,
      "learning_rate": 3.041978148361127e-05,
      "loss": 0.086,
      "step": 7810
    },
    {
      "epoch": 2.1261555193039694,
      "grad_norm": 1.078125,
      "learning_rate": 3.0391029327199544e-05,
      "loss": 0.0936,
      "step": 7820
    },
    {
      "epoch": 2.128874388254486,
      "grad_norm": 1.4296875,
      "learning_rate": 3.036227717078781e-05,
      "loss": 0.1141,
      "step": 7830
    },
    {
      "epoch": 2.1315932572050027,
      "grad_norm": 1.6328125,
      "learning_rate": 3.033352501437608e-05,
      "loss": 0.0887,
      "step": 7840
    },
    {
      "epoch": 2.1343121261555194,
      "grad_norm": 1.578125,
      "learning_rate": 3.0304772857964346e-05,
      "loss": 0.0945,
      "step": 7850
    },
    {
      "epoch": 2.1343121261555194,
      "eval_loss": 0.12286718934774399,
      "eval_runtime": 161.0286,
      "eval_samples_per_second": 3.105,
      "eval_steps_per_second": 0.099,
      "step": 7850
    },
    {
      "epoch": 2.137030995106036,
      "grad_norm": 2.765625,
      "learning_rate": 3.0276020701552617e-05,
      "loss": 0.1248,
      "step": 7860
    },
    {
      "epoch": 2.1397498640565527,
      "grad_norm": 2.328125,
      "learning_rate": 3.0247268545140888e-05,
      "loss": 0.0921,
      "step": 7870
    },
    {
      "epoch": 2.142468733007069,
      "grad_norm": 1.1640625,
      "learning_rate": 3.021851638872916e-05,
      "loss": 0.0907,
      "step": 7880
    },
    {
      "epoch": 2.1451876019575855,
      "grad_norm": 1.4921875,
      "learning_rate": 3.0189764232317426e-05,
      "loss": 0.1019,
      "step": 7890
    },
    {
      "epoch": 2.147906470908102,
      "grad_norm": 1.1875,
      "learning_rate": 3.0161012075905693e-05,
      "loss": 0.101,
      "step": 7900
    },
    {
      "epoch": 2.147906470908102,
      "eval_loss": 0.11469531059265137,
      "eval_runtime": 103.7149,
      "eval_samples_per_second": 4.821,
      "eval_steps_per_second": 0.154,
      "step": 7900
    },
    {
      "epoch": 2.150625339858619,
      "grad_norm": 1.1328125,
      "learning_rate": 3.0132259919493967e-05,
      "loss": 0.1069,
      "step": 7910
    },
    {
      "epoch": 2.1533442088091355,
      "grad_norm": 0.6484375,
      "learning_rate": 3.0103507763082234e-05,
      "loss": 0.0901,
      "step": 7920
    },
    {
      "epoch": 2.156063077759652,
      "grad_norm": 1.359375,
      "learning_rate": 3.00747556066705e-05,
      "loss": 0.1071,
      "step": 7930
    },
    {
      "epoch": 2.1587819467101688,
      "grad_norm": 1.6953125,
      "learning_rate": 3.004600345025877e-05,
      "loss": 0.0848,
      "step": 7940
    },
    {
      "epoch": 2.161500815660685,
      "grad_norm": 1.125,
      "learning_rate": 3.0017251293847036e-05,
      "loss": 0.0941,
      "step": 7950
    },
    {
      "epoch": 2.161500815660685,
      "eval_loss": 0.11516406387090683,
      "eval_runtime": 123.8821,
      "eval_samples_per_second": 4.036,
      "eval_steps_per_second": 0.129,
      "step": 7950
    },
    {
      "epoch": 2.1642196846112016,
      "grad_norm": 1.828125,
      "learning_rate": 2.998849913743531e-05,
      "loss": 0.0989,
      "step": 7960
    },
    {
      "epoch": 2.1669385535617183,
      "grad_norm": 0.76171875,
      "learning_rate": 2.9959746981023578e-05,
      "loss": 0.1012,
      "step": 7970
    },
    {
      "epoch": 2.169657422512235,
      "grad_norm": 1.140625,
      "learning_rate": 2.9930994824611845e-05,
      "loss": 0.1054,
      "step": 7980
    },
    {
      "epoch": 2.1723762914627516,
      "grad_norm": 0.984375,
      "learning_rate": 2.9902242668200116e-05,
      "loss": 0.0949,
      "step": 7990
    },
    {
      "epoch": 2.1750951604132682,
      "grad_norm": 1.1328125,
      "learning_rate": 2.9873490511788386e-05,
      "loss": 0.0986,
      "step": 8000
    },
    {
      "epoch": 2.1750951604132682,
      "eval_loss": 0.11370312422513962,
      "eval_runtime": 154.5113,
      "eval_samples_per_second": 3.236,
      "eval_steps_per_second": 0.104,
      "step": 8000
    },
    {
      "epoch": 2.177814029363785,
      "grad_norm": 2.71875,
      "learning_rate": 2.9844738355376657e-05,
      "loss": 0.1258,
      "step": 8010
    },
    {
      "epoch": 2.180532898314301,
      "grad_norm": 1.890625,
      "learning_rate": 2.9815986198964924e-05,
      "loss": 0.0981,
      "step": 8020
    },
    {
      "epoch": 2.1832517672648177,
      "grad_norm": 0.90234375,
      "learning_rate": 2.9787234042553192e-05,
      "loss": 0.1003,
      "step": 8030
    },
    {
      "epoch": 2.1859706362153344,
      "grad_norm": 2.296875,
      "learning_rate": 2.975848188614146e-05,
      "loss": 0.1272,
      "step": 8040
    },
    {
      "epoch": 2.188689505165851,
      "grad_norm": 1.9921875,
      "learning_rate": 2.9729729729729733e-05,
      "loss": 0.1219,
      "step": 8050
    },
    {
      "epoch": 2.188689505165851,
      "eval_loss": 0.11237499862909317,
      "eval_runtime": 104.3876,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.153,
      "step": 8050
    },
    {
      "epoch": 2.1914083741163677,
      "grad_norm": 2.875,
      "learning_rate": 2.9700977573318e-05,
      "loss": 0.1144,
      "step": 8060
    },
    {
      "epoch": 2.1941272430668843,
      "grad_norm": 1.9453125,
      "learning_rate": 2.9672225416906268e-05,
      "loss": 0.0997,
      "step": 8070
    },
    {
      "epoch": 2.1968461120174005,
      "grad_norm": 0.79296875,
      "learning_rate": 2.9643473260494535e-05,
      "loss": 0.0973,
      "step": 8080
    },
    {
      "epoch": 2.199564980967917,
      "grad_norm": 0.9921875,
      "learning_rate": 2.961472110408281e-05,
      "loss": 0.0953,
      "step": 8090
    },
    {
      "epoch": 2.202283849918434,
      "grad_norm": 1.4609375,
      "learning_rate": 2.9585968947671077e-05,
      "loss": 0.0738,
      "step": 8100
    },
    {
      "epoch": 2.202283849918434,
      "eval_loss": 0.12086718529462814,
      "eval_runtime": 98.221,
      "eval_samples_per_second": 5.091,
      "eval_steps_per_second": 0.163,
      "step": 8100
    },
    {
      "epoch": 2.2050027188689505,
      "grad_norm": 0.71484375,
      "learning_rate": 2.9557216791259347e-05,
      "loss": 0.1036,
      "step": 8110
    },
    {
      "epoch": 2.207721587819467,
      "grad_norm": 1.3359375,
      "learning_rate": 2.9528464634847615e-05,
      "loss": 0.0929,
      "step": 8120
    },
    {
      "epoch": 2.210440456769984,
      "grad_norm": 1.2734375,
      "learning_rate": 2.9499712478435882e-05,
      "loss": 0.0883,
      "step": 8130
    },
    {
      "epoch": 2.2131593257205004,
      "grad_norm": 1.59375,
      "learning_rate": 2.9470960322024156e-05,
      "loss": 0.1233,
      "step": 8140
    },
    {
      "epoch": 2.2158781946710167,
      "grad_norm": 1.078125,
      "learning_rate": 2.9442208165612423e-05,
      "loss": 0.1095,
      "step": 8150
    },
    {
      "epoch": 2.2158781946710167,
      "eval_loss": 0.1126406267285347,
      "eval_runtime": 97.6582,
      "eval_samples_per_second": 5.12,
      "eval_steps_per_second": 0.164,
      "step": 8150
    },
    {
      "epoch": 2.2185970636215333,
      "grad_norm": 2.015625,
      "learning_rate": 2.941345600920069e-05,
      "loss": 0.0939,
      "step": 8160
    },
    {
      "epoch": 2.22131593257205,
      "grad_norm": 1.9140625,
      "learning_rate": 2.9384703852788958e-05,
      "loss": 0.1156,
      "step": 8170
    },
    {
      "epoch": 2.2240348015225666,
      "grad_norm": 1.9609375,
      "learning_rate": 2.9355951696377232e-05,
      "loss": 0.1201,
      "step": 8180
    },
    {
      "epoch": 2.2267536704730833,
      "grad_norm": 1.625,
      "learning_rate": 2.93271995399655e-05,
      "loss": 0.0846,
      "step": 8190
    },
    {
      "epoch": 2.2294725394236,
      "grad_norm": 2.28125,
      "learning_rate": 2.9298447383553767e-05,
      "loss": 0.1057,
      "step": 8200
    },
    {
      "epoch": 2.2294725394236,
      "eval_loss": 0.11463281512260437,
      "eval_runtime": 98.6806,
      "eval_samples_per_second": 5.067,
      "eval_steps_per_second": 0.162,
      "step": 8200
    },
    {
      "epoch": 2.2321914083741166,
      "grad_norm": 1.6875,
      "learning_rate": 2.9269695227142034e-05,
      "loss": 0.1176,
      "step": 8210
    },
    {
      "epoch": 2.2349102773246328,
      "grad_norm": 2.078125,
      "learning_rate": 2.9240943070730308e-05,
      "loss": 0.1022,
      "step": 8220
    },
    {
      "epoch": 2.2376291462751494,
      "grad_norm": 1.640625,
      "learning_rate": 2.9212190914318575e-05,
      "loss": 0.1073,
      "step": 8230
    },
    {
      "epoch": 2.240348015225666,
      "grad_norm": 1.4765625,
      "learning_rate": 2.9183438757906846e-05,
      "loss": 0.1003,
      "step": 8240
    },
    {
      "epoch": 2.2430668841761827,
      "grad_norm": 0.796875,
      "learning_rate": 2.9154686601495113e-05,
      "loss": 0.083,
      "step": 8250
    },
    {
      "epoch": 2.2430668841761827,
      "eval_loss": 0.10762500017881393,
      "eval_runtime": 97.6464,
      "eval_samples_per_second": 5.121,
      "eval_steps_per_second": 0.164,
      "step": 8250
    },
    {
      "epoch": 2.2457857531266994,
      "grad_norm": 1.3359375,
      "learning_rate": 2.912593444508338e-05,
      "loss": 0.088,
      "step": 8260
    },
    {
      "epoch": 2.248504622077216,
      "grad_norm": 1.0546875,
      "learning_rate": 2.9097182288671655e-05,
      "loss": 0.1178,
      "step": 8270
    },
    {
      "epoch": 2.2512234910277327,
      "grad_norm": 1.3984375,
      "learning_rate": 2.9068430132259922e-05,
      "loss": 0.1082,
      "step": 8280
    },
    {
      "epoch": 2.253942359978249,
      "grad_norm": 1.7421875,
      "learning_rate": 2.903967797584819e-05,
      "loss": 0.1123,
      "step": 8290
    },
    {
      "epoch": 2.2566612289287655,
      "grad_norm": 1.0,
      "learning_rate": 2.9010925819436457e-05,
      "loss": 0.0817,
      "step": 8300
    },
    {
      "epoch": 2.2566612289287655,
      "eval_loss": 0.10864843428134918,
      "eval_runtime": 96.8766,
      "eval_samples_per_second": 5.161,
      "eval_steps_per_second": 0.165,
      "step": 8300
    },
    {
      "epoch": 2.259380097879282,
      "grad_norm": 2.234375,
      "learning_rate": 2.898217366302473e-05,
      "loss": 0.1322,
      "step": 8310
    },
    {
      "epoch": 2.262098966829799,
      "grad_norm": 1.2734375,
      "learning_rate": 2.8953421506612998e-05,
      "loss": 0.1117,
      "step": 8320
    },
    {
      "epoch": 2.2648178357803155,
      "grad_norm": 2.71875,
      "learning_rate": 2.8924669350201265e-05,
      "loss": 0.0985,
      "step": 8330
    },
    {
      "epoch": 2.267536704730832,
      "grad_norm": 0.9140625,
      "learning_rate": 2.8895917193789536e-05,
      "loss": 0.0925,
      "step": 8340
    },
    {
      "epoch": 2.2702555736813483,
      "grad_norm": 3.515625,
      "learning_rate": 2.8867165037377803e-05,
      "loss": 0.1158,
      "step": 8350
    },
    {
      "epoch": 2.2702555736813483,
      "eval_loss": 0.11531250178813934,
      "eval_runtime": 109.5485,
      "eval_samples_per_second": 4.564,
      "eval_steps_per_second": 0.146,
      "step": 8350
    },
    {
      "epoch": 2.272974442631865,
      "grad_norm": 2.03125,
      "learning_rate": 2.8838412880966078e-05,
      "loss": 0.1124,
      "step": 8360
    },
    {
      "epoch": 2.2756933115823816,
      "grad_norm": 2.046875,
      "learning_rate": 2.8809660724554345e-05,
      "loss": 0.1118,
      "step": 8370
    },
    {
      "epoch": 2.2784121805328983,
      "grad_norm": 2.125,
      "learning_rate": 2.8780908568142612e-05,
      "loss": 0.0845,
      "step": 8380
    },
    {
      "epoch": 2.281131049483415,
      "grad_norm": 1.421875,
      "learning_rate": 2.875215641173088e-05,
      "loss": 0.1199,
      "step": 8390
    },
    {
      "epoch": 2.2838499184339316,
      "grad_norm": 0.70703125,
      "learning_rate": 2.8723404255319154e-05,
      "loss": 0.092,
      "step": 8400
    },
    {
      "epoch": 2.2838499184339316,
      "eval_loss": 0.11608593910932541,
      "eval_runtime": 103.2061,
      "eval_samples_per_second": 4.845,
      "eval_steps_per_second": 0.155,
      "step": 8400
    },
    {
      "epoch": 2.2865687873844482,
      "grad_norm": 1.671875,
      "learning_rate": 2.869465209890742e-05,
      "loss": 0.1199,
      "step": 8410
    },
    {
      "epoch": 2.289287656334965,
      "grad_norm": 2.046875,
      "learning_rate": 2.8665899942495688e-05,
      "loss": 0.1114,
      "step": 8420
    },
    {
      "epoch": 2.292006525285481,
      "grad_norm": 1.4296875,
      "learning_rate": 2.8637147786083956e-05,
      "loss": 0.1063,
      "step": 8430
    },
    {
      "epoch": 2.2947253942359978,
      "grad_norm": 1.6328125,
      "learning_rate": 2.8608395629672223e-05,
      "loss": 0.1242,
      "step": 8440
    },
    {
      "epoch": 2.2974442631865144,
      "grad_norm": 1.0390625,
      "learning_rate": 2.8579643473260497e-05,
      "loss": 0.099,
      "step": 8450
    },
    {
      "epoch": 2.2974442631865144,
      "eval_loss": 0.1130390614271164,
      "eval_runtime": 98.0839,
      "eval_samples_per_second": 5.098,
      "eval_steps_per_second": 0.163,
      "step": 8450
    },
    {
      "epoch": 2.300163132137031,
      "grad_norm": 1.3671875,
      "learning_rate": 2.8550891316848764e-05,
      "loss": 0.0883,
      "step": 8460
    },
    {
      "epoch": 2.3028820010875477,
      "grad_norm": 1.53125,
      "learning_rate": 2.8522139160437035e-05,
      "loss": 0.0844,
      "step": 8470
    },
    {
      "epoch": 2.3056008700380644,
      "grad_norm": 1.0546875,
      "learning_rate": 2.8493387004025302e-05,
      "loss": 0.0946,
      "step": 8480
    },
    {
      "epoch": 2.3083197389885806,
      "grad_norm": 2.75,
      "learning_rate": 2.8464634847613576e-05,
      "loss": 0.103,
      "step": 8490
    },
    {
      "epoch": 2.311038607939097,
      "grad_norm": 1.3125,
      "learning_rate": 2.8435882691201844e-05,
      "loss": 0.1106,
      "step": 8500
    },
    {
      "epoch": 2.311038607939097,
      "eval_loss": 0.1290312558412552,
      "eval_runtime": 97.3972,
      "eval_samples_per_second": 5.134,
      "eval_steps_per_second": 0.164,
      "step": 8500
    },
    {
      "epoch": 2.313757476889614,
      "grad_norm": 1.4375,
      "learning_rate": 2.840713053479011e-05,
      "loss": 0.1357,
      "step": 8510
    },
    {
      "epoch": 2.3164763458401305,
      "grad_norm": 1.84375,
      "learning_rate": 2.8378378378378378e-05,
      "loss": 0.0863,
      "step": 8520
    },
    {
      "epoch": 2.319195214790647,
      "grad_norm": 2.0,
      "learning_rate": 2.8349626221966646e-05,
      "loss": 0.1139,
      "step": 8530
    },
    {
      "epoch": 2.321914083741164,
      "grad_norm": 1.9609375,
      "learning_rate": 2.832087406555492e-05,
      "loss": 0.0965,
      "step": 8540
    },
    {
      "epoch": 2.3246329526916805,
      "grad_norm": 1.2265625,
      "learning_rate": 2.8292121909143187e-05,
      "loss": 0.1111,
      "step": 8550
    },
    {
      "epoch": 2.3246329526916805,
      "eval_loss": 0.11896874755620956,
      "eval_runtime": 98.5482,
      "eval_samples_per_second": 5.074,
      "eval_steps_per_second": 0.162,
      "step": 8550
    },
    {
      "epoch": 2.3273518216421967,
      "grad_norm": 2.015625,
      "learning_rate": 2.8263369752731454e-05,
      "loss": 0.1137,
      "step": 8560
    },
    {
      "epoch": 2.3300706905927133,
      "grad_norm": 1.2890625,
      "learning_rate": 2.8234617596319725e-05,
      "loss": 0.1182,
      "step": 8570
    },
    {
      "epoch": 2.33278955954323,
      "grad_norm": 2.421875,
      "learning_rate": 2.8205865439907996e-05,
      "loss": 0.1157,
      "step": 8580
    },
    {
      "epoch": 2.3355084284937466,
      "grad_norm": 2.015625,
      "learning_rate": 2.8177113283496266e-05,
      "loss": 0.0936,
      "step": 8590
    },
    {
      "epoch": 2.3382272974442633,
      "grad_norm": 1.3359375,
      "learning_rate": 2.8148361127084534e-05,
      "loss": 0.0715,
      "step": 8600
    },
    {
      "epoch": 2.3382272974442633,
      "eval_loss": 0.1123046875,
      "eval_runtime": 105.1834,
      "eval_samples_per_second": 4.754,
      "eval_steps_per_second": 0.152,
      "step": 8600
    },
    {
      "epoch": 2.34094616639478,
      "grad_norm": 1.984375,
      "learning_rate": 2.81196089706728e-05,
      "loss": 0.1132,
      "step": 8610
    },
    {
      "epoch": 2.343665035345296,
      "grad_norm": 0.6640625,
      "learning_rate": 2.809085681426107e-05,
      "loss": 0.1017,
      "step": 8620
    },
    {
      "epoch": 2.346383904295813,
      "grad_norm": 2.375,
      "learning_rate": 2.8062104657849342e-05,
      "loss": 0.0854,
      "step": 8630
    },
    {
      "epoch": 2.3491027732463294,
      "grad_norm": 4.4375,
      "learning_rate": 2.803335250143761e-05,
      "loss": 0.0955,
      "step": 8640
    },
    {
      "epoch": 2.351821642196846,
      "grad_norm": 2.84375,
      "learning_rate": 2.8004600345025877e-05,
      "loss": 0.1075,
      "step": 8650
    },
    {
      "epoch": 2.351821642196846,
      "eval_loss": 0.12440624833106995,
      "eval_runtime": 96.8148,
      "eval_samples_per_second": 5.165,
      "eval_steps_per_second": 0.165,
      "step": 8650
    },
    {
      "epoch": 2.3545405111473627,
      "grad_norm": 1.328125,
      "learning_rate": 2.7975848188614144e-05,
      "loss": 0.1082,
      "step": 8660
    },
    {
      "epoch": 2.3572593800978794,
      "grad_norm": 1.140625,
      "learning_rate": 2.794709603220242e-05,
      "loss": 0.1022,
      "step": 8670
    },
    {
      "epoch": 2.359978249048396,
      "grad_norm": 0.65625,
      "learning_rate": 2.7918343875790686e-05,
      "loss": 0.105,
      "step": 8680
    },
    {
      "epoch": 2.3626971179989127,
      "grad_norm": 1.2109375,
      "learning_rate": 2.7889591719378953e-05,
      "loss": 0.1136,
      "step": 8690
    },
    {
      "epoch": 2.365415986949429,
      "grad_norm": 2.078125,
      "learning_rate": 2.7860839562967224e-05,
      "loss": 0.1239,
      "step": 8700
    },
    {
      "epoch": 2.365415986949429,
      "eval_loss": 0.12122656404972076,
      "eval_runtime": 101.7762,
      "eval_samples_per_second": 4.913,
      "eval_steps_per_second": 0.157,
      "step": 8700
    },
    {
      "epoch": 2.3681348558999455,
      "grad_norm": 1.5859375,
      "learning_rate": 2.783208740655549e-05,
      "loss": 0.115,
      "step": 8710
    },
    {
      "epoch": 2.370853724850462,
      "grad_norm": 1.2109375,
      "learning_rate": 2.7803335250143765e-05,
      "loss": 0.1003,
      "step": 8720
    },
    {
      "epoch": 2.373572593800979,
      "grad_norm": 1.9765625,
      "learning_rate": 2.7774583093732033e-05,
      "loss": 0.1202,
      "step": 8730
    },
    {
      "epoch": 2.3762914627514955,
      "grad_norm": 1.203125,
      "learning_rate": 2.77458309373203e-05,
      "loss": 0.1351,
      "step": 8740
    },
    {
      "epoch": 2.379010331702012,
      "grad_norm": 1.59375,
      "learning_rate": 2.7717078780908567e-05,
      "loss": 0.0935,
      "step": 8750
    },
    {
      "epoch": 2.379010331702012,
      "eval_loss": 0.12915624678134918,
      "eval_runtime": 95.8182,
      "eval_samples_per_second": 5.218,
      "eval_steps_per_second": 0.167,
      "step": 8750
    },
    {
      "epoch": 2.3817292006525284,
      "grad_norm": 1.5859375,
      "learning_rate": 2.768832662449684e-05,
      "loss": 0.1035,
      "step": 8760
    },
    {
      "epoch": 2.384448069603045,
      "grad_norm": 1.9375,
      "learning_rate": 2.765957446808511e-05,
      "loss": 0.0737,
      "step": 8770
    },
    {
      "epoch": 2.3871669385535617,
      "grad_norm": 1.4375,
      "learning_rate": 2.7630822311673376e-05,
      "loss": 0.0935,
      "step": 8780
    },
    {
      "epoch": 2.3898858075040783,
      "grad_norm": 1.2109375,
      "learning_rate": 2.7602070155261643e-05,
      "loss": 0.0822,
      "step": 8790
    },
    {
      "epoch": 2.392604676454595,
      "grad_norm": 1.7421875,
      "learning_rate": 2.7573317998849914e-05,
      "loss": 0.1,
      "step": 8800
    },
    {
      "epoch": 2.392604676454595,
      "eval_loss": 0.11945312470197678,
      "eval_runtime": 95.5791,
      "eval_samples_per_second": 5.231,
      "eval_steps_per_second": 0.167,
      "step": 8800
    },
    {
      "epoch": 2.3953235454051116,
      "grad_norm": 1.0,
      "learning_rate": 2.7544565842438185e-05,
      "loss": 0.0873,
      "step": 8810
    },
    {
      "epoch": 2.3980424143556283,
      "grad_norm": 1.3359375,
      "learning_rate": 2.7515813686026455e-05,
      "loss": 0.1072,
      "step": 8820
    },
    {
      "epoch": 2.4007612833061445,
      "grad_norm": 2.125,
      "learning_rate": 2.7487061529614723e-05,
      "loss": 0.088,
      "step": 8830
    },
    {
      "epoch": 2.403480152256661,
      "grad_norm": 0.91015625,
      "learning_rate": 2.745830937320299e-05,
      "loss": 0.0923,
      "step": 8840
    },
    {
      "epoch": 2.4061990212071778,
      "grad_norm": 1.015625,
      "learning_rate": 2.7429557216791264e-05,
      "loss": 0.1313,
      "step": 8850
    },
    {
      "epoch": 2.4061990212071778,
      "eval_loss": 0.12060156464576721,
      "eval_runtime": 108.7757,
      "eval_samples_per_second": 4.597,
      "eval_steps_per_second": 0.147,
      "step": 8850
    },
    {
      "epoch": 2.4089178901576944,
      "grad_norm": 2.265625,
      "learning_rate": 2.740080506037953e-05,
      "loss": 0.108,
      "step": 8860
    },
    {
      "epoch": 2.411636759108211,
      "grad_norm": 2.234375,
      "learning_rate": 2.73720529039678e-05,
      "loss": 0.0908,
      "step": 8870
    },
    {
      "epoch": 2.4143556280587277,
      "grad_norm": 0.94140625,
      "learning_rate": 2.7343300747556066e-05,
      "loss": 0.0739,
      "step": 8880
    },
    {
      "epoch": 2.417074497009244,
      "grad_norm": 2.078125,
      "learning_rate": 2.7314548591144333e-05,
      "loss": 0.1102,
      "step": 8890
    },
    {
      "epoch": 2.4197933659597606,
      "grad_norm": 1.265625,
      "learning_rate": 2.7285796434732607e-05,
      "loss": 0.0893,
      "step": 8900
    },
    {
      "epoch": 2.4197933659597606,
      "eval_loss": 0.12201562523841858,
      "eval_runtime": 99.5365,
      "eval_samples_per_second": 5.023,
      "eval_steps_per_second": 0.161,
      "step": 8900
    },
    {
      "epoch": 2.4225122349102772,
      "grad_norm": 1.265625,
      "learning_rate": 2.7257044278320875e-05,
      "loss": 0.083,
      "step": 8910
    },
    {
      "epoch": 2.425231103860794,
      "grad_norm": 3.390625,
      "learning_rate": 2.7228292121909142e-05,
      "loss": 0.0933,
      "step": 8920
    },
    {
      "epoch": 2.4279499728113105,
      "grad_norm": 0.84375,
      "learning_rate": 2.7199539965497413e-05,
      "loss": 0.1052,
      "step": 8930
    },
    {
      "epoch": 2.430668841761827,
      "grad_norm": 1.765625,
      "learning_rate": 2.7170787809085683e-05,
      "loss": 0.0833,
      "step": 8940
    },
    {
      "epoch": 2.433387710712344,
      "grad_norm": 1.75,
      "learning_rate": 2.7142035652673954e-05,
      "loss": 0.0901,
      "step": 8950
    },
    {
      "epoch": 2.433387710712344,
      "eval_loss": 0.12113281339406967,
      "eval_runtime": 100.0255,
      "eval_samples_per_second": 4.999,
      "eval_steps_per_second": 0.16,
      "step": 8950
    },
    {
      "epoch": 2.4361065796628605,
      "grad_norm": 1.734375,
      "learning_rate": 2.711328349626222e-05,
      "loss": 0.1049,
      "step": 8960
    },
    {
      "epoch": 2.4388254486133767,
      "grad_norm": 1.6953125,
      "learning_rate": 2.708453133985049e-05,
      "loss": 0.1075,
      "step": 8970
    },
    {
      "epoch": 2.4415443175638933,
      "grad_norm": 1.9921875,
      "learning_rate": 2.7055779183438763e-05,
      "loss": 0.1028,
      "step": 8980
    },
    {
      "epoch": 2.44426318651441,
      "grad_norm": 1.5390625,
      "learning_rate": 2.702702702702703e-05,
      "loss": 0.0937,
      "step": 8990
    },
    {
      "epoch": 2.4469820554649266,
      "grad_norm": 2.234375,
      "learning_rate": 2.6998274870615297e-05,
      "loss": 0.1041,
      "step": 9000
    },
    {
      "epoch": 2.4469820554649266,
      "eval_loss": 0.11175781488418579,
      "eval_runtime": 100.2909,
      "eval_samples_per_second": 4.985,
      "eval_steps_per_second": 0.16,
      "step": 9000
    },
    {
      "epoch": 2.4497009244154433,
      "grad_norm": 1.328125,
      "learning_rate": 2.6969522714203565e-05,
      "loss": 0.0909,
      "step": 9010
    },
    {
      "epoch": 2.45241979336596,
      "grad_norm": 2.203125,
      "learning_rate": 2.6940770557791832e-05,
      "loss": 0.0964,
      "step": 9020
    },
    {
      "epoch": 2.455138662316476,
      "grad_norm": 1.4375,
      "learning_rate": 2.6912018401380106e-05,
      "loss": 0.0831,
      "step": 9030
    },
    {
      "epoch": 2.457857531266993,
      "grad_norm": 1.375,
      "learning_rate": 2.6883266244968374e-05,
      "loss": 0.1061,
      "step": 9040
    },
    {
      "epoch": 2.4605764002175095,
      "grad_norm": 2.734375,
      "learning_rate": 2.6854514088556644e-05,
      "loss": 0.1046,
      "step": 9050
    },
    {
      "epoch": 2.4605764002175095,
      "eval_loss": 0.11450780928134918,
      "eval_runtime": 99.2852,
      "eval_samples_per_second": 5.036,
      "eval_steps_per_second": 0.161,
      "step": 9050
    },
    {
      "epoch": 2.463295269168026,
      "grad_norm": 1.328125,
      "learning_rate": 2.682576193214491e-05,
      "loss": 0.1078,
      "step": 9060
    },
    {
      "epoch": 2.4660141381185428,
      "grad_norm": 1.703125,
      "learning_rate": 2.6797009775733186e-05,
      "loss": 0.133,
      "step": 9070
    },
    {
      "epoch": 2.4687330070690594,
      "grad_norm": 1.1796875,
      "learning_rate": 2.6768257619321453e-05,
      "loss": 0.0752,
      "step": 9080
    },
    {
      "epoch": 2.471451876019576,
      "grad_norm": 1.90625,
      "learning_rate": 2.673950546290972e-05,
      "loss": 0.1011,
      "step": 9090
    },
    {
      "epoch": 2.4741707449700923,
      "grad_norm": 1.921875,
      "learning_rate": 2.6710753306497988e-05,
      "loss": 0.1112,
      "step": 9100
    },
    {
      "epoch": 2.4741707449700923,
      "eval_loss": 0.10976562649011612,
      "eval_runtime": 132.045,
      "eval_samples_per_second": 3.787,
      "eval_steps_per_second": 0.121,
      "step": 9100
    },
    {
      "epoch": 2.476889613920609,
      "grad_norm": 1.8671875,
      "learning_rate": 2.6682001150086255e-05,
      "loss": 0.1048,
      "step": 9110
    },
    {
      "epoch": 2.4796084828711256,
      "grad_norm": 1.46875,
      "learning_rate": 2.665324899367453e-05,
      "loss": 0.1285,
      "step": 9120
    },
    {
      "epoch": 2.482327351821642,
      "grad_norm": 1.2109375,
      "learning_rate": 2.6624496837262796e-05,
      "loss": 0.0892,
      "step": 9130
    },
    {
      "epoch": 2.485046220772159,
      "grad_norm": 1.28125,
      "learning_rate": 2.6595744680851064e-05,
      "loss": 0.0926,
      "step": 9140
    },
    {
      "epoch": 2.4877650897226755,
      "grad_norm": 1.5859375,
      "learning_rate": 2.656699252443933e-05,
      "loss": 0.0905,
      "step": 9150
    },
    {
      "epoch": 2.4877650897226755,
      "eval_loss": 0.11485937237739563,
      "eval_runtime": 96.7607,
      "eval_samples_per_second": 5.167,
      "eval_steps_per_second": 0.165,
      "step": 9150
    },
    {
      "epoch": 2.4904839586731917,
      "grad_norm": 2.984375,
      "learning_rate": 2.6538240368027605e-05,
      "loss": 0.1083,
      "step": 9160
    },
    {
      "epoch": 2.4932028276237084,
      "grad_norm": 2.21875,
      "learning_rate": 2.6509488211615872e-05,
      "loss": 0.1062,
      "step": 9170
    },
    {
      "epoch": 2.495921696574225,
      "grad_norm": 1.1171875,
      "learning_rate": 2.6480736055204143e-05,
      "loss": 0.086,
      "step": 9180
    },
    {
      "epoch": 2.4986405655247417,
      "grad_norm": 1.9453125,
      "learning_rate": 2.645198389879241e-05,
      "loss": 0.1208,
      "step": 9190
    },
    {
      "epoch": 2.5013594344752583,
      "grad_norm": 1.171875,
      "learning_rate": 2.6423231742380678e-05,
      "loss": 0.0905,
      "step": 9200
    },
    {
      "epoch": 2.5013594344752583,
      "eval_loss": 0.12715624272823334,
      "eval_runtime": 98.5321,
      "eval_samples_per_second": 5.074,
      "eval_steps_per_second": 0.162,
      "step": 9200
    },
    {
      "epoch": 2.504078303425775,
      "grad_norm": 1.3828125,
      "learning_rate": 2.6394479585968952e-05,
      "loss": 0.1231,
      "step": 9210
    },
    {
      "epoch": 2.5067971723762916,
      "grad_norm": 2.703125,
      "learning_rate": 2.636572742955722e-05,
      "loss": 0.1287,
      "step": 9220
    },
    {
      "epoch": 2.5095160413268083,
      "grad_norm": 1.2421875,
      "learning_rate": 2.6336975273145486e-05,
      "loss": 0.1048,
      "step": 9230
    },
    {
      "epoch": 2.5122349102773245,
      "grad_norm": 1.4921875,
      "learning_rate": 2.6308223116733754e-05,
      "loss": 0.0927,
      "step": 9240
    },
    {
      "epoch": 2.514953779227841,
      "grad_norm": 2.34375,
      "learning_rate": 2.6279470960322028e-05,
      "loss": 0.1211,
      "step": 9250
    },
    {
      "epoch": 2.514953779227841,
      "eval_loss": 0.11287499964237213,
      "eval_runtime": 96.8468,
      "eval_samples_per_second": 5.163,
      "eval_steps_per_second": 0.165,
      "step": 9250
    },
    {
      "epoch": 2.517672648178358,
      "grad_norm": 3.671875,
      "learning_rate": 2.6250718803910295e-05,
      "loss": 0.1128,
      "step": 9260
    },
    {
      "epoch": 2.5203915171288744,
      "grad_norm": 1.4609375,
      "learning_rate": 2.6221966647498562e-05,
      "loss": 0.0896,
      "step": 9270
    },
    {
      "epoch": 2.523110386079391,
      "grad_norm": 2.4375,
      "learning_rate": 2.6193214491086833e-05,
      "loss": 0.0987,
      "step": 9280
    },
    {
      "epoch": 2.5258292550299073,
      "grad_norm": 2.3125,
      "learning_rate": 2.61644623346751e-05,
      "loss": 0.1062,
      "step": 9290
    },
    {
      "epoch": 2.528548123980424,
      "grad_norm": 1.78125,
      "learning_rate": 2.6135710178263375e-05,
      "loss": 0.0972,
      "step": 9300
    },
    {
      "epoch": 2.528548123980424,
      "eval_loss": 0.12171874940395355,
      "eval_runtime": 96.0517,
      "eval_samples_per_second": 5.206,
      "eval_steps_per_second": 0.167,
      "step": 9300
    },
    {
      "epoch": 2.5312669929309406,
      "grad_norm": 2.40625,
      "learning_rate": 2.6106958021851642e-05,
      "loss": 0.0923,
      "step": 9310
    },
    {
      "epoch": 2.5339858618814572,
      "grad_norm": 1.921875,
      "learning_rate": 2.607820586543991e-05,
      "loss": 0.1005,
      "step": 9320
    },
    {
      "epoch": 2.536704730831974,
      "grad_norm": 1.59375,
      "learning_rate": 2.6049453709028176e-05,
      "loss": 0.1022,
      "step": 9330
    },
    {
      "epoch": 2.5394235997824905,
      "grad_norm": 0.57421875,
      "learning_rate": 2.602070155261645e-05,
      "loss": 0.1033,
      "step": 9340
    },
    {
      "epoch": 2.542142468733007,
      "grad_norm": 2.34375,
      "learning_rate": 2.5991949396204718e-05,
      "loss": 0.0924,
      "step": 9350
    },
    {
      "epoch": 2.542142468733007,
      "eval_loss": 0.12832030653953552,
      "eval_runtime": 116.0797,
      "eval_samples_per_second": 4.307,
      "eval_steps_per_second": 0.138,
      "step": 9350
    },
    {
      "epoch": 2.544861337683524,
      "grad_norm": 1.7109375,
      "learning_rate": 2.5963197239792985e-05,
      "loss": 0.1012,
      "step": 9360
    },
    {
      "epoch": 2.5475802066340405,
      "grad_norm": 1.0234375,
      "learning_rate": 2.5934445083381252e-05,
      "loss": 0.108,
      "step": 9370
    },
    {
      "epoch": 2.5502990755845567,
      "grad_norm": 1.4765625,
      "learning_rate": 2.590569292696952e-05,
      "loss": 0.1038,
      "step": 9380
    },
    {
      "epoch": 2.5530179445350734,
      "grad_norm": 0.66796875,
      "learning_rate": 2.5876940770557794e-05,
      "loss": 0.0886,
      "step": 9390
    },
    {
      "epoch": 2.55573681348559,
      "grad_norm": 1.0859375,
      "learning_rate": 2.584818861414606e-05,
      "loss": 0.0926,
      "step": 9400
    },
    {
      "epoch": 2.55573681348559,
      "eval_loss": 0.11837500333786011,
      "eval_runtime": 105.9318,
      "eval_samples_per_second": 4.72,
      "eval_steps_per_second": 0.151,
      "step": 9400
    },
    {
      "epoch": 2.5584556824361067,
      "grad_norm": 1.1484375,
      "learning_rate": 2.5819436457734332e-05,
      "loss": 0.101,
      "step": 9410
    },
    {
      "epoch": 2.5611745513866233,
      "grad_norm": 1.0703125,
      "learning_rate": 2.57906843013226e-05,
      "loss": 0.0824,
      "step": 9420
    },
    {
      "epoch": 2.5638934203371395,
      "grad_norm": 2.09375,
      "learning_rate": 2.5761932144910873e-05,
      "loss": 0.1157,
      "step": 9430
    },
    {
      "epoch": 2.566612289287656,
      "grad_norm": 2.484375,
      "learning_rate": 2.573317998849914e-05,
      "loss": 0.093,
      "step": 9440
    },
    {
      "epoch": 2.569331158238173,
      "grad_norm": 0.9453125,
      "learning_rate": 2.5704427832087408e-05,
      "loss": 0.1339,
      "step": 9450
    },
    {
      "epoch": 2.569331158238173,
      "eval_loss": 0.13410937786102295,
      "eval_runtime": 101.6961,
      "eval_samples_per_second": 4.917,
      "eval_steps_per_second": 0.157,
      "step": 9450
    },
    {
      "epoch": 2.5720500271886895,
      "grad_norm": 1.53125,
      "learning_rate": 2.5675675675675675e-05,
      "loss": 0.0882,
      "step": 9460
    },
    {
      "epoch": 2.574768896139206,
      "grad_norm": 1.421875,
      "learning_rate": 2.5646923519263943e-05,
      "loss": 0.1068,
      "step": 9470
    },
    {
      "epoch": 2.5774877650897228,
      "grad_norm": 1.859375,
      "learning_rate": 2.5618171362852217e-05,
      "loss": 0.0932,
      "step": 9480
    },
    {
      "epoch": 2.5802066340402394,
      "grad_norm": 1.6875,
      "learning_rate": 2.5589419206440484e-05,
      "loss": 0.0916,
      "step": 9490
    },
    {
      "epoch": 2.582925502990756,
      "grad_norm": 2.890625,
      "learning_rate": 2.556066705002875e-05,
      "loss": 0.1024,
      "step": 9500
    },
    {
      "epoch": 2.582925502990756,
      "eval_loss": 0.13469530642032623,
      "eval_runtime": 98.9064,
      "eval_samples_per_second": 5.055,
      "eval_steps_per_second": 0.162,
      "step": 9500
    },
    {
      "epoch": 2.5856443719412723,
      "grad_norm": 1.609375,
      "learning_rate": 2.5531914893617022e-05,
      "loss": 0.0945,
      "step": 9510
    },
    {
      "epoch": 2.588363240891789,
      "grad_norm": 1.015625,
      "learning_rate": 2.5503162737205293e-05,
      "loss": 0.1088,
      "step": 9520
    },
    {
      "epoch": 2.5910821098423056,
      "grad_norm": 1.921875,
      "learning_rate": 2.5474410580793563e-05,
      "loss": 0.1027,
      "step": 9530
    },
    {
      "epoch": 2.5938009787928222,
      "grad_norm": 1.1875,
      "learning_rate": 2.544565842438183e-05,
      "loss": 0.1035,
      "step": 9540
    },
    {
      "epoch": 2.596519847743339,
      "grad_norm": 1.8515625,
      "learning_rate": 2.5416906267970098e-05,
      "loss": 0.1182,
      "step": 9550
    },
    {
      "epoch": 2.596519847743339,
      "eval_loss": 0.11936718970537186,
      "eval_runtime": 99.204,
      "eval_samples_per_second": 5.04,
      "eval_steps_per_second": 0.161,
      "step": 9550
    },
    {
      "epoch": 2.599238716693855,
      "grad_norm": 0.921875,
      "learning_rate": 2.5388154111558365e-05,
      "loss": 0.1164,
      "step": 9560
    },
    {
      "epoch": 2.6019575856443717,
      "grad_norm": 1.7109375,
      "learning_rate": 2.535940195514664e-05,
      "loss": 0.101,
      "step": 9570
    },
    {
      "epoch": 2.6046764545948884,
      "grad_norm": 2.171875,
      "learning_rate": 2.5330649798734907e-05,
      "loss": 0.0974,
      "step": 9580
    },
    {
      "epoch": 2.607395323545405,
      "grad_norm": 3.671875,
      "learning_rate": 2.5301897642323174e-05,
      "loss": 0.1008,
      "step": 9590
    },
    {
      "epoch": 2.6101141924959217,
      "grad_norm": 1.171875,
      "learning_rate": 2.527314548591144e-05,
      "loss": 0.126,
      "step": 9600
    },
    {
      "epoch": 2.6101141924959217,
      "eval_loss": 0.12602344155311584,
      "eval_runtime": 98.1473,
      "eval_samples_per_second": 5.094,
      "eval_steps_per_second": 0.163,
      "step": 9600
    },
    {
      "epoch": 2.6128330614464383,
      "grad_norm": 0.765625,
      "learning_rate": 2.5244393329499715e-05,
      "loss": 0.1104,
      "step": 9610
    },
    {
      "epoch": 2.615551930396955,
      "grad_norm": 2.046875,
      "learning_rate": 2.5215641173087983e-05,
      "loss": 0.0979,
      "step": 9620
    },
    {
      "epoch": 2.6182707993474716,
      "grad_norm": 1.40625,
      "learning_rate": 2.518688901667625e-05,
      "loss": 0.1086,
      "step": 9630
    },
    {
      "epoch": 2.6209896682979883,
      "grad_norm": 1.265625,
      "learning_rate": 2.515813686026452e-05,
      "loss": 0.093,
      "step": 9640
    },
    {
      "epoch": 2.6237085372485045,
      "grad_norm": 1.9765625,
      "learning_rate": 2.5129384703852788e-05,
      "loss": 0.1031,
      "step": 9650
    },
    {
      "epoch": 2.6237085372485045,
      "eval_loss": 0.11711718887090683,
      "eval_runtime": 98.2211,
      "eval_samples_per_second": 5.091,
      "eval_steps_per_second": 0.163,
      "step": 9650
    },
    {
      "epoch": 2.626427406199021,
      "grad_norm": 1.9765625,
      "learning_rate": 2.5100632547441062e-05,
      "loss": 0.1031,
      "step": 9660
    },
    {
      "epoch": 2.629146275149538,
      "grad_norm": 1.8359375,
      "learning_rate": 2.507188039102933e-05,
      "loss": 0.1093,
      "step": 9670
    },
    {
      "epoch": 2.6318651441000545,
      "grad_norm": 1.9375,
      "learning_rate": 2.5043128234617597e-05,
      "loss": 0.0872,
      "step": 9680
    },
    {
      "epoch": 2.634584013050571,
      "grad_norm": 1.640625,
      "learning_rate": 2.5014376078205864e-05,
      "loss": 0.1013,
      "step": 9690
    },
    {
      "epoch": 2.6373028820010873,
      "grad_norm": 1.25,
      "learning_rate": 2.4985623921794135e-05,
      "loss": 0.0977,
      "step": 9700
    },
    {
      "epoch": 2.6373028820010873,
      "eval_loss": 0.1123671904206276,
      "eval_runtime": 96.8972,
      "eval_samples_per_second": 5.16,
      "eval_steps_per_second": 0.165,
      "step": 9700
    },
    {
      "epoch": 2.640021750951604,
      "grad_norm": 2.328125,
      "learning_rate": 2.4956871765382406e-05,
      "loss": 0.0948,
      "step": 9710
    },
    {
      "epoch": 2.6427406199021206,
      "grad_norm": 2.125,
      "learning_rate": 2.4928119608970673e-05,
      "loss": 0.0881,
      "step": 9720
    },
    {
      "epoch": 2.6454594888526373,
      "grad_norm": 2.359375,
      "learning_rate": 2.4899367452558944e-05,
      "loss": 0.0872,
      "step": 9730
    },
    {
      "epoch": 2.648178357803154,
      "grad_norm": 1.4296875,
      "learning_rate": 2.487061529614721e-05,
      "loss": 0.0975,
      "step": 9740
    },
    {
      "epoch": 2.6508972267536706,
      "grad_norm": 1.171875,
      "learning_rate": 2.484186313973548e-05,
      "loss": 0.1055,
      "step": 9750
    },
    {
      "epoch": 2.6508972267536706,
      "eval_loss": 0.11472656577825546,
      "eval_runtime": 97.6684,
      "eval_samples_per_second": 5.119,
      "eval_steps_per_second": 0.164,
      "step": 9750
    },
    {
      "epoch": 2.653616095704187,
      "grad_norm": 1.5234375,
      "learning_rate": 2.4813110983323752e-05,
      "loss": 0.0875,
      "step": 9760
    },
    {
      "epoch": 2.656334964654704,
      "grad_norm": 0.984375,
      "learning_rate": 2.478435882691202e-05,
      "loss": 0.0943,
      "step": 9770
    },
    {
      "epoch": 2.65905383360522,
      "grad_norm": 1.0625,
      "learning_rate": 2.475560667050029e-05,
      "loss": 0.0876,
      "step": 9780
    },
    {
      "epoch": 2.6617727025557367,
      "grad_norm": 1.6875,
      "learning_rate": 2.4726854514088558e-05,
      "loss": 0.0958,
      "step": 9790
    },
    {
      "epoch": 2.6644915715062534,
      "grad_norm": 2.734375,
      "learning_rate": 2.469810235767683e-05,
      "loss": 0.117,
      "step": 9800
    },
    {
      "epoch": 2.6644915715062534,
      "eval_loss": 0.12026562541723251,
      "eval_runtime": 96.4718,
      "eval_samples_per_second": 5.183,
      "eval_steps_per_second": 0.166,
      "step": 9800
    },
    {
      "epoch": 2.66721044045677,
      "grad_norm": 1.5234375,
      "learning_rate": 2.4669350201265096e-05,
      "loss": 0.116,
      "step": 9810
    },
    {
      "epoch": 2.6699293094072867,
      "grad_norm": 1.1328125,
      "learning_rate": 2.4640598044853366e-05,
      "loss": 0.0971,
      "step": 9820
    },
    {
      "epoch": 2.672648178357803,
      "grad_norm": 1.84375,
      "learning_rate": 2.4611845888441634e-05,
      "loss": 0.1266,
      "step": 9830
    },
    {
      "epoch": 2.6753670473083195,
      "grad_norm": 1.0859375,
      "learning_rate": 2.45830937320299e-05,
      "loss": 0.096,
      "step": 9840
    },
    {
      "epoch": 2.678085916258836,
      "grad_norm": 2.671875,
      "learning_rate": 2.455434157561817e-05,
      "loss": 0.1055,
      "step": 9850
    },
    {
      "epoch": 2.678085916258836,
      "eval_loss": 0.11753906309604645,
      "eval_runtime": 98.2289,
      "eval_samples_per_second": 5.09,
      "eval_steps_per_second": 0.163,
      "step": 9850
    },
    {
      "epoch": 2.680804785209353,
      "grad_norm": 1.6953125,
      "learning_rate": 2.452558941920644e-05,
      "loss": 0.1174,
      "step": 9860
    },
    {
      "epoch": 2.6835236541598695,
      "grad_norm": 1.078125,
      "learning_rate": 2.449683726279471e-05,
      "loss": 0.1307,
      "step": 9870
    },
    {
      "epoch": 2.686242523110386,
      "grad_norm": 0.90625,
      "learning_rate": 2.446808510638298e-05,
      "loss": 0.1097,
      "step": 9880
    },
    {
      "epoch": 2.688961392060903,
      "grad_norm": 2.421875,
      "learning_rate": 2.443933294997125e-05,
      "loss": 0.1083,
      "step": 9890
    },
    {
      "epoch": 2.6916802610114194,
      "grad_norm": 1.3359375,
      "learning_rate": 2.441058079355952e-05,
      "loss": 0.0943,
      "step": 9900
    },
    {
      "epoch": 2.6916802610114194,
      "eval_loss": 0.13484375178813934,
      "eval_runtime": 117.0805,
      "eval_samples_per_second": 4.271,
      "eval_steps_per_second": 0.137,
      "step": 9900
    },
    {
      "epoch": 2.694399129961936,
      "grad_norm": 1.3671875,
      "learning_rate": 2.438182863714779e-05,
      "loss": 0.0974,
      "step": 9910
    },
    {
      "epoch": 2.6971179989124523,
      "grad_norm": 1.1953125,
      "learning_rate": 2.4353076480736056e-05,
      "loss": 0.131,
      "step": 9920
    },
    {
      "epoch": 2.699836867862969,
      "grad_norm": 1.875,
      "learning_rate": 2.4324324324324327e-05,
      "loss": 0.0895,
      "step": 9930
    },
    {
      "epoch": 2.7025557368134856,
      "grad_norm": 7.375,
      "learning_rate": 2.4295572167912594e-05,
      "loss": 0.1071,
      "step": 9940
    },
    {
      "epoch": 2.7052746057640022,
      "grad_norm": 1.328125,
      "learning_rate": 2.4266820011500862e-05,
      "loss": 0.1008,
      "step": 9950
    },
    {
      "epoch": 2.7052746057640022,
      "eval_loss": 0.11382031440734863,
      "eval_runtime": 102.6495,
      "eval_samples_per_second": 4.871,
      "eval_steps_per_second": 0.156,
      "step": 9950
    },
    {
      "epoch": 2.707993474714519,
      "grad_norm": 1.6796875,
      "learning_rate": 2.4238067855089132e-05,
      "loss": 0.1077,
      "step": 9960
    },
    {
      "epoch": 2.710712343665035,
      "grad_norm": 1.53125,
      "learning_rate": 2.42093156986774e-05,
      "loss": 0.1115,
      "step": 9970
    },
    {
      "epoch": 2.7134312126155518,
      "grad_norm": 2.53125,
      "learning_rate": 2.418056354226567e-05,
      "loss": 0.1029,
      "step": 9980
    },
    {
      "epoch": 2.7161500815660684,
      "grad_norm": 0.423828125,
      "learning_rate": 2.415181138585394e-05,
      "loss": 0.0834,
      "step": 9990
    },
    {
      "epoch": 2.718868950516585,
      "grad_norm": 1.390625,
      "learning_rate": 2.4123059229442212e-05,
      "loss": 0.1074,
      "step": 10000
    },
    {
      "epoch": 2.718868950516585,
      "eval_loss": 0.12809374928474426,
      "eval_runtime": 103.0236,
      "eval_samples_per_second": 4.853,
      "eval_steps_per_second": 0.155,
      "step": 10000
    },
    {
      "epoch": 2.7215878194671017,
      "grad_norm": 3.40625,
      "learning_rate": 2.409430707303048e-05,
      "loss": 0.1094,
      "step": 10010
    },
    {
      "epoch": 2.7243066884176184,
      "grad_norm": 1.2109375,
      "learning_rate": 2.406555491661875e-05,
      "loss": 0.0893,
      "step": 10020
    },
    {
      "epoch": 2.727025557368135,
      "grad_norm": 1.4296875,
      "learning_rate": 2.4036802760207017e-05,
      "loss": 0.0911,
      "step": 10030
    },
    {
      "epoch": 2.7297444263186517,
      "grad_norm": 2.203125,
      "learning_rate": 2.4008050603795285e-05,
      "loss": 0.1017,
      "step": 10040
    },
    {
      "epoch": 2.732463295269168,
      "grad_norm": 1.03125,
      "learning_rate": 2.3979298447383555e-05,
      "loss": 0.1093,
      "step": 10050
    },
    {
      "epoch": 2.732463295269168,
      "eval_loss": 0.12435156106948853,
      "eval_runtime": 101.1794,
      "eval_samples_per_second": 4.942,
      "eval_steps_per_second": 0.158,
      "step": 10050
    },
    {
      "epoch": 2.7351821642196845,
      "grad_norm": 1.8125,
      "learning_rate": 2.3950546290971823e-05,
      "loss": 0.0762,
      "step": 10060
    },
    {
      "epoch": 2.737901033170201,
      "grad_norm": 1.5625,
      "learning_rate": 2.3921794134560093e-05,
      "loss": 0.0809,
      "step": 10070
    },
    {
      "epoch": 2.740619902120718,
      "grad_norm": 1.6171875,
      "learning_rate": 2.389304197814836e-05,
      "loss": 0.1073,
      "step": 10080
    },
    {
      "epoch": 2.7433387710712345,
      "grad_norm": 1.2578125,
      "learning_rate": 2.386428982173663e-05,
      "loss": 0.0984,
      "step": 10090
    },
    {
      "epoch": 2.7460576400217507,
      "grad_norm": 0.9609375,
      "learning_rate": 2.38355376653249e-05,
      "loss": 0.1096,
      "step": 10100
    },
    {
      "epoch": 2.7460576400217507,
      "eval_loss": 0.12822656333446503,
      "eval_runtime": 100.451,
      "eval_samples_per_second": 4.978,
      "eval_steps_per_second": 0.159,
      "step": 10100
    },
    {
      "epoch": 2.7487765089722673,
      "grad_norm": 0.8046875,
      "learning_rate": 2.380678550891317e-05,
      "loss": 0.0812,
      "step": 10110
    },
    {
      "epoch": 2.751495377922784,
      "grad_norm": 3.015625,
      "learning_rate": 2.377803335250144e-05,
      "loss": 0.1125,
      "step": 10120
    },
    {
      "epoch": 2.7542142468733006,
      "grad_norm": 1.7734375,
      "learning_rate": 2.3749281196089707e-05,
      "loss": 0.0964,
      "step": 10130
    },
    {
      "epoch": 2.7569331158238173,
      "grad_norm": 1.5078125,
      "learning_rate": 2.3720529039677978e-05,
      "loss": 0.106,
      "step": 10140
    },
    {
      "epoch": 2.759651984774334,
      "grad_norm": 2.078125,
      "learning_rate": 2.3691776883266245e-05,
      "loss": 0.1035,
      "step": 10150
    },
    {
      "epoch": 2.759651984774334,
      "eval_loss": 0.11612500250339508,
      "eval_runtime": 101.786,
      "eval_samples_per_second": 4.912,
      "eval_steps_per_second": 0.157,
      "step": 10150
    },
    {
      "epoch": 2.7623708537248506,
      "grad_norm": 1.640625,
      "learning_rate": 2.3663024726854516e-05,
      "loss": 0.0816,
      "step": 10160
    },
    {
      "epoch": 2.7650897226753672,
      "grad_norm": 1.421875,
      "learning_rate": 2.3634272570442783e-05,
      "loss": 0.0915,
      "step": 10170
    },
    {
      "epoch": 2.767808591625884,
      "grad_norm": 2.015625,
      "learning_rate": 2.3605520414031054e-05,
      "loss": 0.1122,
      "step": 10180
    },
    {
      "epoch": 2.7705274605764,
      "grad_norm": 1.328125,
      "learning_rate": 2.357676825761932e-05,
      "loss": 0.0845,
      "step": 10190
    },
    {
      "epoch": 2.7732463295269167,
      "grad_norm": 2.28125,
      "learning_rate": 2.3548016101207592e-05,
      "loss": 0.1123,
      "step": 10200
    },
    {
      "epoch": 2.7732463295269167,
      "eval_loss": 0.12323437631130219,
      "eval_runtime": 96.6642,
      "eval_samples_per_second": 5.173,
      "eval_steps_per_second": 0.166,
      "step": 10200
    },
    {
      "epoch": 2.7759651984774334,
      "grad_norm": 1.578125,
      "learning_rate": 2.351926394479586e-05,
      "loss": 0.1263,
      "step": 10210
    },
    {
      "epoch": 2.77868406742795,
      "grad_norm": 1.7890625,
      "learning_rate": 2.349051178838413e-05,
      "loss": 0.0929,
      "step": 10220
    },
    {
      "epoch": 2.7814029363784667,
      "grad_norm": 2.265625,
      "learning_rate": 2.34617596319724e-05,
      "loss": 0.0943,
      "step": 10230
    },
    {
      "epoch": 2.784121805328983,
      "grad_norm": 2.078125,
      "learning_rate": 2.3433007475560668e-05,
      "loss": 0.0744,
      "step": 10240
    },
    {
      "epoch": 2.7868406742794996,
      "grad_norm": 1.734375,
      "learning_rate": 2.340425531914894e-05,
      "loss": 0.1007,
      "step": 10250
    },
    {
      "epoch": 2.7868406742794996,
      "eval_loss": 0.10999219119548798,
      "eval_runtime": 99.029,
      "eval_samples_per_second": 5.049,
      "eval_steps_per_second": 0.162,
      "step": 10250
    },
    {
      "epoch": 2.789559543230016,
      "grad_norm": 1.7109375,
      "learning_rate": 2.3375503162737206e-05,
      "loss": 0.0939,
      "step": 10260
    },
    {
      "epoch": 2.792278412180533,
      "grad_norm": 1.046875,
      "learning_rate": 2.3346751006325477e-05,
      "loss": 0.0982,
      "step": 10270
    },
    {
      "epoch": 2.7949972811310495,
      "grad_norm": 1.71875,
      "learning_rate": 2.3317998849913744e-05,
      "loss": 0.0984,
      "step": 10280
    },
    {
      "epoch": 2.797716150081566,
      "grad_norm": 1.5390625,
      "learning_rate": 2.3289246693502015e-05,
      "loss": 0.1067,
      "step": 10290
    },
    {
      "epoch": 2.800435019032083,
      "grad_norm": 1.2421875,
      "learning_rate": 2.3260494537090282e-05,
      "loss": 0.0695,
      "step": 10300
    },
    {
      "epoch": 2.800435019032083,
      "eval_loss": 0.12232030928134918,
      "eval_runtime": 96.6067,
      "eval_samples_per_second": 5.176,
      "eval_steps_per_second": 0.166,
      "step": 10300
    },
    {
      "epoch": 2.8031538879825995,
      "grad_norm": 1.421875,
      "learning_rate": 2.3231742380678553e-05,
      "loss": 0.0887,
      "step": 10310
    },
    {
      "epoch": 2.8058727569331157,
      "grad_norm": 2.546875,
      "learning_rate": 2.320299022426682e-05,
      "loss": 0.0895,
      "step": 10320
    },
    {
      "epoch": 2.8085916258836323,
      "grad_norm": 1.3125,
      "learning_rate": 2.3174238067855087e-05,
      "loss": 0.1028,
      "step": 10330
    },
    {
      "epoch": 2.811310494834149,
      "grad_norm": 2.609375,
      "learning_rate": 2.3145485911443358e-05,
      "loss": 0.1052,
      "step": 10340
    },
    {
      "epoch": 2.8140293637846656,
      "grad_norm": 2.109375,
      "learning_rate": 2.311673375503163e-05,
      "loss": 0.1084,
      "step": 10350
    },
    {
      "epoch": 2.8140293637846656,
      "eval_loss": 0.11949999630451202,
      "eval_runtime": 98.1142,
      "eval_samples_per_second": 5.096,
      "eval_steps_per_second": 0.163,
      "step": 10350
    },
    {
      "epoch": 2.8167482327351823,
      "grad_norm": 1.703125,
      "learning_rate": 2.30879815986199e-05,
      "loss": 0.1153,
      "step": 10360
    },
    {
      "epoch": 2.8194671016856985,
      "grad_norm": 1.6015625,
      "learning_rate": 2.3059229442208167e-05,
      "loss": 0.0856,
      "step": 10370
    },
    {
      "epoch": 2.822185970636215,
      "grad_norm": 1.3671875,
      "learning_rate": 2.3030477285796438e-05,
      "loss": 0.0979,
      "step": 10380
    },
    {
      "epoch": 2.8249048395867318,
      "grad_norm": 1.4140625,
      "learning_rate": 2.3001725129384705e-05,
      "loss": 0.0991,
      "step": 10390
    },
    {
      "epoch": 2.8276237085372484,
      "grad_norm": 1.4609375,
      "learning_rate": 2.2972972972972976e-05,
      "loss": 0.0948,
      "step": 10400
    },
    {
      "epoch": 2.8276237085372484,
      "eval_loss": 0.10865624994039536,
      "eval_runtime": 111.7934,
      "eval_samples_per_second": 4.473,
      "eval_steps_per_second": 0.143,
      "step": 10400
    },
    {
      "epoch": 2.830342577487765,
      "grad_norm": 2.0625,
      "learning_rate": 2.2944220816561243e-05,
      "loss": 0.0648,
      "step": 10410
    },
    {
      "epoch": 2.8330614464382817,
      "grad_norm": 1.59375,
      "learning_rate": 2.291546866014951e-05,
      "loss": 0.1039,
      "step": 10420
    },
    {
      "epoch": 2.8357803153887984,
      "grad_norm": 0.95703125,
      "learning_rate": 2.288671650373778e-05,
      "loss": 0.1145,
      "step": 10430
    },
    {
      "epoch": 2.838499184339315,
      "grad_norm": 1.25,
      "learning_rate": 2.2857964347326048e-05,
      "loss": 0.0835,
      "step": 10440
    },
    {
      "epoch": 2.8412180532898317,
      "grad_norm": 1.4140625,
      "learning_rate": 2.282921219091432e-05,
      "loss": 0.1037,
      "step": 10450
    },
    {
      "epoch": 2.8412180532898317,
      "eval_loss": 0.11501562595367432,
      "eval_runtime": 100.906,
      "eval_samples_per_second": 4.955,
      "eval_steps_per_second": 0.159,
      "step": 10450
    },
    {
      "epoch": 2.843936922240348,
      "grad_norm": 0.58984375,
      "learning_rate": 2.280046003450259e-05,
      "loss": 0.0712,
      "step": 10460
    },
    {
      "epoch": 2.8466557911908645,
      "grad_norm": 4.25,
      "learning_rate": 2.277170787809086e-05,
      "loss": 0.102,
      "step": 10470
    },
    {
      "epoch": 2.849374660141381,
      "grad_norm": 1.125,
      "learning_rate": 2.2742955721679128e-05,
      "loss": 0.0751,
      "step": 10480
    },
    {
      "epoch": 2.852093529091898,
      "grad_norm": 2.078125,
      "learning_rate": 2.27142035652674e-05,
      "loss": 0.1099,
      "step": 10490
    },
    {
      "epoch": 2.8548123980424145,
      "grad_norm": 0.65234375,
      "learning_rate": 2.2685451408855666e-05,
      "loss": 0.0893,
      "step": 10500
    },
    {
      "epoch": 2.8548123980424145,
      "eval_loss": 0.11222656071186066,
      "eval_runtime": 99.4228,
      "eval_samples_per_second": 5.029,
      "eval_steps_per_second": 0.161,
      "step": 10500
    },
    {
      "epoch": 2.8575312669929307,
      "grad_norm": 1.078125,
      "learning_rate": 2.2656699252443933e-05,
      "loss": 0.109,
      "step": 10510
    },
    {
      "epoch": 2.8602501359434473,
      "grad_norm": 2.75,
      "learning_rate": 2.2627947096032204e-05,
      "loss": 0.0928,
      "step": 10520
    },
    {
      "epoch": 2.862969004893964,
      "grad_norm": 0.6796875,
      "learning_rate": 2.259919493962047e-05,
      "loss": 0.0981,
      "step": 10530
    },
    {
      "epoch": 2.8656878738444806,
      "grad_norm": 2.375,
      "learning_rate": 2.2570442783208742e-05,
      "loss": 0.1156,
      "step": 10540
    },
    {
      "epoch": 2.8684067427949973,
      "grad_norm": 2.453125,
      "learning_rate": 2.254169062679701e-05,
      "loss": 0.1106,
      "step": 10550
    },
    {
      "epoch": 2.8684067427949973,
      "eval_loss": 0.12965625524520874,
      "eval_runtime": 98.6922,
      "eval_samples_per_second": 5.066,
      "eval_steps_per_second": 0.162,
      "step": 10550
    },
    {
      "epoch": 2.871125611745514,
      "grad_norm": 1.4921875,
      "learning_rate": 2.251293847038528e-05,
      "loss": 0.1129,
      "step": 10560
    },
    {
      "epoch": 2.8738444806960306,
      "grad_norm": 2.140625,
      "learning_rate": 2.2484186313973547e-05,
      "loss": 0.0931,
      "step": 10570
    },
    {
      "epoch": 2.8765633496465473,
      "grad_norm": 1.390625,
      "learning_rate": 2.2455434157561818e-05,
      "loss": 0.0931,
      "step": 10580
    },
    {
      "epoch": 2.8792822185970635,
      "grad_norm": 1.921875,
      "learning_rate": 2.242668200115009e-05,
      "loss": 0.0844,
      "step": 10590
    },
    {
      "epoch": 2.88200108754758,
      "grad_norm": 1.7890625,
      "learning_rate": 2.2397929844738356e-05,
      "loss": 0.0918,
      "step": 10600
    },
    {
      "epoch": 2.88200108754758,
      "eval_loss": 0.11282812803983688,
      "eval_runtime": 97.1529,
      "eval_samples_per_second": 5.147,
      "eval_steps_per_second": 0.165,
      "step": 10600
    },
    {
      "epoch": 2.8847199564980968,
      "grad_norm": 3.125,
      "learning_rate": 2.2369177688326627e-05,
      "loss": 0.1001,
      "step": 10610
    },
    {
      "epoch": 2.8874388254486134,
      "grad_norm": 2.46875,
      "learning_rate": 2.2340425531914894e-05,
      "loss": 0.102,
      "step": 10620
    },
    {
      "epoch": 2.89015769439913,
      "grad_norm": 1.6015625,
      "learning_rate": 2.2311673375503165e-05,
      "loss": 0.1039,
      "step": 10630
    },
    {
      "epoch": 2.8928765633496467,
      "grad_norm": 2.4375,
      "learning_rate": 2.2282921219091432e-05,
      "loss": 0.101,
      "step": 10640
    },
    {
      "epoch": 2.895595432300163,
      "grad_norm": 2.109375,
      "learning_rate": 2.2254169062679703e-05,
      "loss": 0.097,
      "step": 10650
    },
    {
      "epoch": 2.895595432300163,
      "eval_loss": 0.12345312535762787,
      "eval_runtime": 99.1892,
      "eval_samples_per_second": 5.041,
      "eval_steps_per_second": 0.161,
      "step": 10650
    },
    {
      "epoch": 2.8983143012506796,
      "grad_norm": 2.4375,
      "learning_rate": 2.222541690626797e-05,
      "loss": 0.1039,
      "step": 10660
    },
    {
      "epoch": 2.901033170201196,
      "grad_norm": 0.81640625,
      "learning_rate": 2.219666474985624e-05,
      "loss": 0.0946,
      "step": 10670
    },
    {
      "epoch": 2.903752039151713,
      "grad_norm": 1.8125,
      "learning_rate": 2.2167912593444508e-05,
      "loss": 0.1042,
      "step": 10680
    },
    {
      "epoch": 2.9064709081022295,
      "grad_norm": 2.1875,
      "learning_rate": 2.213916043703278e-05,
      "loss": 0.093,
      "step": 10690
    },
    {
      "epoch": 2.909189777052746,
      "grad_norm": 0.8125,
      "learning_rate": 2.211040828062105e-05,
      "loss": 0.1049,
      "step": 10700
    },
    {
      "epoch": 2.909189777052746,
      "eval_loss": 0.1251562535762787,
      "eval_runtime": 102.827,
      "eval_samples_per_second": 4.863,
      "eval_steps_per_second": 0.156,
      "step": 10700
    },
    {
      "epoch": 2.911908646003263,
      "grad_norm": 1.2734375,
      "learning_rate": 2.2081656124209317e-05,
      "loss": 0.0905,
      "step": 10710
    },
    {
      "epoch": 2.9146275149537795,
      "grad_norm": 2.953125,
      "learning_rate": 2.2052903967797587e-05,
      "loss": 0.1006,
      "step": 10720
    },
    {
      "epoch": 2.9173463839042957,
      "grad_norm": 0.8203125,
      "learning_rate": 2.2024151811385855e-05,
      "loss": 0.0891,
      "step": 10730
    },
    {
      "epoch": 2.9200652528548123,
      "grad_norm": 2.5625,
      "learning_rate": 2.1995399654974125e-05,
      "loss": 0.0913,
      "step": 10740
    },
    {
      "epoch": 2.922784121805329,
      "grad_norm": 1.328125,
      "learning_rate": 2.1966647498562393e-05,
      "loss": 0.1224,
      "step": 10750
    },
    {
      "epoch": 2.922784121805329,
      "eval_loss": 0.12055468559265137,
      "eval_runtime": 98.0251,
      "eval_samples_per_second": 5.101,
      "eval_steps_per_second": 0.163,
      "step": 10750
    },
    {
      "epoch": 2.9255029907558456,
      "grad_norm": 2.109375,
      "learning_rate": 2.1937895342150663e-05,
      "loss": 0.1135,
      "step": 10760
    },
    {
      "epoch": 2.9282218597063623,
      "grad_norm": 2.28125,
      "learning_rate": 2.190914318573893e-05,
      "loss": 0.11,
      "step": 10770
    },
    {
      "epoch": 2.9309407286568785,
      "grad_norm": 1.9609375,
      "learning_rate": 2.18803910293272e-05,
      "loss": 0.092,
      "step": 10780
    },
    {
      "epoch": 2.933659597607395,
      "grad_norm": 2.21875,
      "learning_rate": 2.185163887291547e-05,
      "loss": 0.1066,
      "step": 10790
    },
    {
      "epoch": 2.936378466557912,
      "grad_norm": 1.90625,
      "learning_rate": 2.1822886716503736e-05,
      "loss": 0.1329,
      "step": 10800
    },
    {
      "epoch": 2.936378466557912,
      "eval_loss": 0.10992187261581421,
      "eval_runtime": 99.0811,
      "eval_samples_per_second": 5.046,
      "eval_steps_per_second": 0.161,
      "step": 10800
    },
    {
      "epoch": 2.9390973355084284,
      "grad_norm": 0.921875,
      "learning_rate": 2.1794134560092007e-05,
      "loss": 0.0789,
      "step": 10810
    },
    {
      "epoch": 2.941816204458945,
      "grad_norm": 1.3125,
      "learning_rate": 2.1765382403680277e-05,
      "loss": 0.0879,
      "step": 10820
    },
    {
      "epoch": 2.9445350734094617,
      "grad_norm": 1.8125,
      "learning_rate": 2.1736630247268548e-05,
      "loss": 0.1066,
      "step": 10830
    },
    {
      "epoch": 2.9472539423599784,
      "grad_norm": 3.9375,
      "learning_rate": 2.1707878090856815e-05,
      "loss": 0.1082,
      "step": 10840
    },
    {
      "epoch": 2.949972811310495,
      "grad_norm": 1.3671875,
      "learning_rate": 2.1679125934445086e-05,
      "loss": 0.0775,
      "step": 10850
    },
    {
      "epoch": 2.949972811310495,
      "eval_loss": 0.11228124797344208,
      "eval_runtime": 96.5815,
      "eval_samples_per_second": 5.177,
      "eval_steps_per_second": 0.166,
      "step": 10850
    },
    {
      "epoch": 2.9526916802610113,
      "grad_norm": 0.99609375,
      "learning_rate": 2.1650373778033353e-05,
      "loss": 0.0822,
      "step": 10860
    },
    {
      "epoch": 2.955410549211528,
      "grad_norm": 2.34375,
      "learning_rate": 2.1621621621621624e-05,
      "loss": 0.1136,
      "step": 10870
    },
    {
      "epoch": 2.9581294181620446,
      "grad_norm": 1.4765625,
      "learning_rate": 2.159286946520989e-05,
      "loss": 0.1104,
      "step": 10880
    },
    {
      "epoch": 2.960848287112561,
      "grad_norm": 1.1484375,
      "learning_rate": 2.156411730879816e-05,
      "loss": 0.1021,
      "step": 10890
    },
    {
      "epoch": 2.963567156063078,
      "grad_norm": 1.8359375,
      "learning_rate": 2.153536515238643e-05,
      "loss": 0.103,
      "step": 10900
    },
    {
      "epoch": 2.963567156063078,
      "eval_loss": 0.11634375154972076,
      "eval_runtime": 112.3213,
      "eval_samples_per_second": 4.452,
      "eval_steps_per_second": 0.142,
      "step": 10900
    },
    {
      "epoch": 2.9662860250135945,
      "grad_norm": 1.4453125,
      "learning_rate": 2.1506612995974697e-05,
      "loss": 0.1039,
      "step": 10910
    },
    {
      "epoch": 2.9690048939641107,
      "grad_norm": 1.5625,
      "learning_rate": 2.1477860839562967e-05,
      "loss": 0.087,
      "step": 10920
    },
    {
      "epoch": 2.9717237629146274,
      "grad_norm": 2.140625,
      "learning_rate": 2.1449108683151238e-05,
      "loss": 0.0906,
      "step": 10930
    },
    {
      "epoch": 2.974442631865144,
      "grad_norm": 2.234375,
      "learning_rate": 2.142035652673951e-05,
      "loss": 0.1104,
      "step": 10940
    },
    {
      "epoch": 2.9771615008156607,
      "grad_norm": 2.515625,
      "learning_rate": 2.1391604370327776e-05,
      "loss": 0.1075,
      "step": 10950
    },
    {
      "epoch": 2.9771615008156607,
      "eval_loss": 0.11387500166893005,
      "eval_runtime": 105.7422,
      "eval_samples_per_second": 4.728,
      "eval_steps_per_second": 0.151,
      "step": 10950
    },
    {
      "epoch": 2.9798803697661773,
      "grad_norm": 2.015625,
      "learning_rate": 2.1362852213916047e-05,
      "loss": 0.1141,
      "step": 10960
    },
    {
      "epoch": 2.982599238716694,
      "grad_norm": 0.73046875,
      "learning_rate": 2.1334100057504314e-05,
      "loss": 0.0706,
      "step": 10970
    },
    {
      "epoch": 2.9853181076672106,
      "grad_norm": 1.34375,
      "learning_rate": 2.130534790109258e-05,
      "loss": 0.1146,
      "step": 10980
    },
    {
      "epoch": 2.9880369766177273,
      "grad_norm": 2.234375,
      "learning_rate": 2.1276595744680852e-05,
      "loss": 0.1054,
      "step": 10990
    },
    {
      "epoch": 2.9907558455682435,
      "grad_norm": 2.390625,
      "learning_rate": 2.124784358826912e-05,
      "loss": 0.1038,
      "step": 11000
    },
    {
      "epoch": 2.9907558455682435,
      "eval_loss": 0.11911718547344208,
      "eval_runtime": 100.2747,
      "eval_samples_per_second": 4.986,
      "eval_steps_per_second": 0.16,
      "step": 11000
    },
    {
      "epoch": 2.99347471451876,
      "grad_norm": 0.7578125,
      "learning_rate": 2.121909143185739e-05,
      "loss": 0.1043,
      "step": 11010
    },
    {
      "epoch": 2.9961935834692768,
      "grad_norm": 1.515625,
      "learning_rate": 2.1190339275445658e-05,
      "loss": 0.1113,
      "step": 11020
    },
    {
      "epoch": 2.9989124524197934,
      "grad_norm": 1.328125,
      "learning_rate": 2.1161587119033928e-05,
      "loss": 0.0899,
      "step": 11030
    },
    {
      "epoch": 3.00163132137031,
      "grad_norm": 1.4765625,
      "learning_rate": 2.1132834962622196e-05,
      "loss": 0.0936,
      "step": 11040
    },
    {
      "epoch": 3.0043501903208267,
      "grad_norm": 1.84375,
      "learning_rate": 2.1104082806210466e-05,
      "loss": 0.0964,
      "step": 11050
    },
    {
      "epoch": 3.0043501903208267,
      "eval_loss": 0.11122656613588333,
      "eval_runtime": 132.3893,
      "eval_samples_per_second": 3.777,
      "eval_steps_per_second": 0.121,
      "step": 11050
    },
    {
      "epoch": 3.007069059271343,
      "grad_norm": 1.4375,
      "learning_rate": 2.1075330649798737e-05,
      "loss": 0.076,
      "step": 11060
    },
    {
      "epoch": 3.0097879282218596,
      "grad_norm": 1.703125,
      "learning_rate": 2.1046578493387008e-05,
      "loss": 0.0851,
      "step": 11070
    },
    {
      "epoch": 3.0125067971723762,
      "grad_norm": 0.9609375,
      "learning_rate": 2.1017826336975275e-05,
      "loss": 0.0794,
      "step": 11080
    },
    {
      "epoch": 3.015225666122893,
      "grad_norm": 0.8203125,
      "learning_rate": 2.0989074180563542e-05,
      "loss": 0.0964,
      "step": 11090
    },
    {
      "epoch": 3.0179445350734095,
      "grad_norm": 0.71875,
      "learning_rate": 2.0960322024151813e-05,
      "loss": 0.0906,
      "step": 11100
    },
    {
      "epoch": 3.0179445350734095,
      "eval_loss": 0.11168749630451202,
      "eval_runtime": 106.7498,
      "eval_samples_per_second": 4.684,
      "eval_steps_per_second": 0.15,
      "step": 11100
    },
    {
      "epoch": 3.020663404023926,
      "grad_norm": 2.3125,
      "learning_rate": 2.093156986774008e-05,
      "loss": 0.0856,
      "step": 11110
    },
    {
      "epoch": 3.023382272974443,
      "grad_norm": 2.90625,
      "learning_rate": 2.090281771132835e-05,
      "loss": 0.103,
      "step": 11120
    },
    {
      "epoch": 3.026101141924959,
      "grad_norm": 1.5,
      "learning_rate": 2.087406555491662e-05,
      "loss": 0.1014,
      "step": 11130
    },
    {
      "epoch": 3.0288200108754757,
      "grad_norm": 2.25,
      "learning_rate": 2.084531339850489e-05,
      "loss": 0.0968,
      "step": 11140
    },
    {
      "epoch": 3.0315388798259923,
      "grad_norm": 3.203125,
      "learning_rate": 2.0816561242093156e-05,
      "loss": 0.0878,
      "step": 11150
    },
    {
      "epoch": 3.0315388798259923,
      "eval_loss": 0.11215624958276749,
      "eval_runtime": 106.5186,
      "eval_samples_per_second": 4.694,
      "eval_steps_per_second": 0.15,
      "step": 11150
    },
    {
      "epoch": 3.034257748776509,
      "grad_norm": 1.375,
      "learning_rate": 2.0787809085681427e-05,
      "loss": 0.0833,
      "step": 11160
    },
    {
      "epoch": 3.0369766177270257,
      "grad_norm": 0.80859375,
      "learning_rate": 2.0759056929269698e-05,
      "loss": 0.0756,
      "step": 11170
    },
    {
      "epoch": 3.0396954866775423,
      "grad_norm": 1.875,
      "learning_rate": 2.0730304772857965e-05,
      "loss": 0.0931,
      "step": 11180
    },
    {
      "epoch": 3.0424143556280585,
      "grad_norm": 2.25,
      "learning_rate": 2.0701552616446236e-05,
      "loss": 0.1086,
      "step": 11190
    },
    {
      "epoch": 3.045133224578575,
      "grad_norm": 1.46875,
      "learning_rate": 2.0672800460034503e-05,
      "loss": 0.101,
      "step": 11200
    },
    {
      "epoch": 3.045133224578575,
      "eval_loss": 0.11080468446016312,
      "eval_runtime": 105.6707,
      "eval_samples_per_second": 4.732,
      "eval_steps_per_second": 0.151,
      "step": 11200
    },
    {
      "epoch": 3.047852093529092,
      "grad_norm": 1.0859375,
      "learning_rate": 2.0644048303622774e-05,
      "loss": 0.0743,
      "step": 11210
    },
    {
      "epoch": 3.0505709624796085,
      "grad_norm": 2.78125,
      "learning_rate": 2.061529614721104e-05,
      "loss": 0.088,
      "step": 11220
    },
    {
      "epoch": 3.053289831430125,
      "grad_norm": 2.03125,
      "learning_rate": 2.0586543990799312e-05,
      "loss": 0.0882,
      "step": 11230
    },
    {
      "epoch": 3.0560087003806418,
      "grad_norm": 0.380859375,
      "learning_rate": 2.055779183438758e-05,
      "loss": 0.0691,
      "step": 11240
    },
    {
      "epoch": 3.0587275693311584,
      "grad_norm": 1.7109375,
      "learning_rate": 2.052903967797585e-05,
      "loss": 0.0862,
      "step": 11250
    },
    {
      "epoch": 3.0587275693311584,
      "eval_loss": 0.11726562678813934,
      "eval_runtime": 102.146,
      "eval_samples_per_second": 4.895,
      "eval_steps_per_second": 0.157,
      "step": 11250
    },
    {
      "epoch": 3.0614464382816746,
      "grad_norm": 2.078125,
      "learning_rate": 2.0500287521564117e-05,
      "loss": 0.081,
      "step": 11260
    },
    {
      "epoch": 3.0641653072321913,
      "grad_norm": 1.6015625,
      "learning_rate": 2.0471535365152384e-05,
      "loss": 0.1009,
      "step": 11270
    },
    {
      "epoch": 3.066884176182708,
      "grad_norm": 0.99609375,
      "learning_rate": 2.0442783208740655e-05,
      "loss": 0.0685,
      "step": 11280
    },
    {
      "epoch": 3.0696030451332246,
      "grad_norm": 2.15625,
      "learning_rate": 2.0414031052328926e-05,
      "loss": 0.0791,
      "step": 11290
    },
    {
      "epoch": 3.0723219140837412,
      "grad_norm": 1.9609375,
      "learning_rate": 2.0385278895917197e-05,
      "loss": 0.0829,
      "step": 11300
    },
    {
      "epoch": 3.0723219140837412,
      "eval_loss": 0.11717187613248825,
      "eval_runtime": 102.3753,
      "eval_samples_per_second": 4.884,
      "eval_steps_per_second": 0.156,
      "step": 11300
    },
    {
      "epoch": 3.075040783034258,
      "grad_norm": 0.76953125,
      "learning_rate": 2.0356526739505464e-05,
      "loss": 0.0927,
      "step": 11310
    },
    {
      "epoch": 3.0777596519847745,
      "grad_norm": 1.75,
      "learning_rate": 2.0327774583093735e-05,
      "loss": 0.0836,
      "step": 11320
    },
    {
      "epoch": 3.0804785209352907,
      "grad_norm": 0.97265625,
      "learning_rate": 2.0299022426682002e-05,
      "loss": 0.0948,
      "step": 11330
    },
    {
      "epoch": 3.0831973898858074,
      "grad_norm": 1.8359375,
      "learning_rate": 2.0270270270270273e-05,
      "loss": 0.1,
      "step": 11340
    },
    {
      "epoch": 3.085916258836324,
      "grad_norm": 1.8984375,
      "learning_rate": 2.024151811385854e-05,
      "loss": 0.0846,
      "step": 11350
    },
    {
      "epoch": 3.085916258836324,
      "eval_loss": 0.1171562522649765,
      "eval_runtime": 132.5537,
      "eval_samples_per_second": 3.772,
      "eval_steps_per_second": 0.121,
      "step": 11350
    },
    {
      "epoch": 3.0886351277868407,
      "grad_norm": 1.515625,
      "learning_rate": 2.0212765957446807e-05,
      "loss": 0.0982,
      "step": 11360
    },
    {
      "epoch": 3.0913539967373573,
      "grad_norm": 2.21875,
      "learning_rate": 2.0184013801035078e-05,
      "loss": 0.0954,
      "step": 11370
    },
    {
      "epoch": 3.094072865687874,
      "grad_norm": 1.140625,
      "learning_rate": 2.0155261644623345e-05,
      "loss": 0.0856,
      "step": 11380
    },
    {
      "epoch": 3.0967917346383906,
      "grad_norm": 0.83984375,
      "learning_rate": 2.0126509488211616e-05,
      "loss": 0.0717,
      "step": 11390
    },
    {
      "epoch": 3.099510603588907,
      "grad_norm": 0.84765625,
      "learning_rate": 2.0097757331799887e-05,
      "loss": 0.0778,
      "step": 11400
    },
    {
      "epoch": 3.099510603588907,
      "eval_loss": 0.12672656774520874,
      "eval_runtime": 108.2612,
      "eval_samples_per_second": 4.618,
      "eval_steps_per_second": 0.148,
      "step": 11400
    },
    {
      "epoch": 3.1022294725394235,
      "grad_norm": 2.75,
      "learning_rate": 2.0069005175388157e-05,
      "loss": 0.0905,
      "step": 11410
    },
    {
      "epoch": 3.10494834148994,
      "grad_norm": 2.0625,
      "learning_rate": 2.0040253018976425e-05,
      "loss": 0.0876,
      "step": 11420
    },
    {
      "epoch": 3.107667210440457,
      "grad_norm": 2.421875,
      "learning_rate": 2.0011500862564695e-05,
      "loss": 0.1259,
      "step": 11430
    },
    {
      "epoch": 3.1103860793909734,
      "grad_norm": 1.6328125,
      "learning_rate": 1.9982748706152963e-05,
      "loss": 0.1008,
      "step": 11440
    },
    {
      "epoch": 3.11310494834149,
      "grad_norm": 2.5625,
      "learning_rate": 1.9953996549741233e-05,
      "loss": 0.0901,
      "step": 11450
    },
    {
      "epoch": 3.11310494834149,
      "eval_loss": 0.11551562696695328,
      "eval_runtime": 106.4481,
      "eval_samples_per_second": 4.697,
      "eval_steps_per_second": 0.15,
      "step": 11450
    },
    {
      "epoch": 3.1158238172920063,
      "grad_norm": 1.359375,
      "learning_rate": 1.99252443933295e-05,
      "loss": 0.1002,
      "step": 11460
    },
    {
      "epoch": 3.118542686242523,
      "grad_norm": 2.40625,
      "learning_rate": 1.9896492236917768e-05,
      "loss": 0.1002,
      "step": 11470
    },
    {
      "epoch": 3.1212615551930396,
      "grad_norm": 1.8359375,
      "learning_rate": 1.986774008050604e-05,
      "loss": 0.0966,
      "step": 11480
    },
    {
      "epoch": 3.1239804241435563,
      "grad_norm": 2.421875,
      "learning_rate": 1.9838987924094306e-05,
      "loss": 0.0924,
      "step": 11490
    },
    {
      "epoch": 3.126699293094073,
      "grad_norm": 4.40625,
      "learning_rate": 1.9810235767682577e-05,
      "loss": 0.0838,
      "step": 11500
    },
    {
      "epoch": 3.126699293094073,
      "eval_loss": 0.11030469089746475,
      "eval_runtime": 104.3763,
      "eval_samples_per_second": 4.79,
      "eval_steps_per_second": 0.153,
      "step": 11500
    },
    {
      "epoch": 3.1294181620445896,
      "grad_norm": 1.53125,
      "learning_rate": 1.9781483611270844e-05,
      "loss": 0.0912,
      "step": 11510
    },
    {
      "epoch": 3.132137030995106,
      "grad_norm": 2.5,
      "learning_rate": 1.9752731454859115e-05,
      "loss": 0.0853,
      "step": 11520
    },
    {
      "epoch": 3.1348558999456224,
      "grad_norm": 1.34375,
      "learning_rate": 1.9723979298447385e-05,
      "loss": 0.0824,
      "step": 11530
    },
    {
      "epoch": 3.137574768896139,
      "grad_norm": 1.5546875,
      "learning_rate": 1.9695227142035656e-05,
      "loss": 0.1096,
      "step": 11540
    },
    {
      "epoch": 3.1402936378466557,
      "grad_norm": 2.09375,
      "learning_rate": 1.9666474985623923e-05,
      "loss": 0.0768,
      "step": 11550
    },
    {
      "epoch": 3.1402936378466557,
      "eval_loss": 0.11176562309265137,
      "eval_runtime": 103.2385,
      "eval_samples_per_second": 4.843,
      "eval_steps_per_second": 0.155,
      "step": 11550
    },
    {
      "epoch": 3.1430125067971724,
      "grad_norm": 2.15625,
      "learning_rate": 1.963772282921219e-05,
      "loss": 0.08,
      "step": 11560
    },
    {
      "epoch": 3.145731375747689,
      "grad_norm": 1.8125,
      "learning_rate": 1.960897067280046e-05,
      "loss": 0.0957,
      "step": 11570
    },
    {
      "epoch": 3.1484502446982057,
      "grad_norm": 2.015625,
      "learning_rate": 1.958021851638873e-05,
      "loss": 0.0685,
      "step": 11580
    },
    {
      "epoch": 3.1511691136487223,
      "grad_norm": 2.0625,
      "learning_rate": 1.9551466359977e-05,
      "loss": 0.0942,
      "step": 11590
    },
    {
      "epoch": 3.1538879825992385,
      "grad_norm": 0.59375,
      "learning_rate": 1.9522714203565267e-05,
      "loss": 0.0909,
      "step": 11600
    },
    {
      "epoch": 3.1538879825992385,
      "eval_loss": 0.12700000405311584,
      "eval_runtime": 103.0932,
      "eval_samples_per_second": 4.85,
      "eval_steps_per_second": 0.155,
      "step": 11600
    },
    {
      "epoch": 3.156606851549755,
      "grad_norm": 2.84375,
      "learning_rate": 1.9493962047153538e-05,
      "loss": 0.0791,
      "step": 11610
    },
    {
      "epoch": 3.159325720500272,
      "grad_norm": 2.046875,
      "learning_rate": 1.9465209890741805e-05,
      "loss": 0.0991,
      "step": 11620
    },
    {
      "epoch": 3.1620445894507885,
      "grad_norm": 2.296875,
      "learning_rate": 1.9436457734330076e-05,
      "loss": 0.0887,
      "step": 11630
    },
    {
      "epoch": 3.164763458401305,
      "grad_norm": 1.328125,
      "learning_rate": 1.9407705577918346e-05,
      "loss": 0.0898,
      "step": 11640
    },
    {
      "epoch": 3.167482327351822,
      "grad_norm": 2.359375,
      "learning_rate": 1.9378953421506614e-05,
      "loss": 0.1034,
      "step": 11650
    },
    {
      "epoch": 3.167482327351822,
      "eval_loss": 0.11140625178813934,
      "eval_runtime": 102.1651,
      "eval_samples_per_second": 4.894,
      "eval_steps_per_second": 0.157,
      "step": 11650
    },
    {
      "epoch": 3.1702011963023384,
      "grad_norm": 0.984375,
      "learning_rate": 1.9350201265094884e-05,
      "loss": 0.0653,
      "step": 11660
    },
    {
      "epoch": 3.1729200652528546,
      "grad_norm": 1.953125,
      "learning_rate": 1.932144910868315e-05,
      "loss": 0.0915,
      "step": 11670
    },
    {
      "epoch": 3.1756389342033713,
      "grad_norm": 1.7578125,
      "learning_rate": 1.9292696952271422e-05,
      "loss": 0.0968,
      "step": 11680
    },
    {
      "epoch": 3.178357803153888,
      "grad_norm": 1.3203125,
      "learning_rate": 1.926394479585969e-05,
      "loss": 0.1042,
      "step": 11690
    },
    {
      "epoch": 3.1810766721044046,
      "grad_norm": 1.0859375,
      "learning_rate": 1.923519263944796e-05,
      "loss": 0.0916,
      "step": 11700
    },
    {
      "epoch": 3.1810766721044046,
      "eval_loss": 0.12619531154632568,
      "eval_runtime": 101.7258,
      "eval_samples_per_second": 4.915,
      "eval_steps_per_second": 0.157,
      "step": 11700
    },
    {
      "epoch": 3.1837955410549212,
      "grad_norm": 0.90625,
      "learning_rate": 1.9206440483036228e-05,
      "loss": 0.0744,
      "step": 11710
    },
    {
      "epoch": 3.186514410005438,
      "grad_norm": 2.828125,
      "learning_rate": 1.91776883266245e-05,
      "loss": 0.0949,
      "step": 11720
    },
    {
      "epoch": 3.189233278955954,
      "grad_norm": 2.453125,
      "learning_rate": 1.9148936170212766e-05,
      "loss": 0.0807,
      "step": 11730
    },
    {
      "epoch": 3.1919521479064707,
      "grad_norm": 1.8125,
      "learning_rate": 1.9120184013801033e-05,
      "loss": 0.0807,
      "step": 11740
    },
    {
      "epoch": 3.1946710168569874,
      "grad_norm": 4.15625,
      "learning_rate": 1.9091431857389304e-05,
      "loss": 0.0832,
      "step": 11750
    },
    {
      "epoch": 3.1946710168569874,
      "eval_loss": 0.11633593589067459,
      "eval_runtime": 101.2244,
      "eval_samples_per_second": 4.94,
      "eval_steps_per_second": 0.158,
      "step": 11750
    },
    {
      "epoch": 3.197389885807504,
      "grad_norm": 1.7890625,
      "learning_rate": 1.9062679700977574e-05,
      "loss": 0.1026,
      "step": 11760
    },
    {
      "epoch": 3.2001087547580207,
      "grad_norm": 1.46875,
      "learning_rate": 1.9033927544565845e-05,
      "loss": 0.0843,
      "step": 11770
    },
    {
      "epoch": 3.2028276237085374,
      "grad_norm": 0.734375,
      "learning_rate": 1.9005175388154112e-05,
      "loss": 0.0767,
      "step": 11780
    },
    {
      "epoch": 3.205546492659054,
      "grad_norm": 1.5078125,
      "learning_rate": 1.8976423231742383e-05,
      "loss": 0.0952,
      "step": 11790
    },
    {
      "epoch": 3.2082653616095707,
      "grad_norm": 1.03125,
      "learning_rate": 1.894767107533065e-05,
      "loss": 0.0725,
      "step": 11800
    },
    {
      "epoch": 3.2082653616095707,
      "eval_loss": 0.11535156518220901,
      "eval_runtime": 100.0738,
      "eval_samples_per_second": 4.996,
      "eval_steps_per_second": 0.16,
      "step": 11800
    },
    {
      "epoch": 3.210984230560087,
      "grad_norm": 1.640625,
      "learning_rate": 1.891891891891892e-05,
      "loss": 0.0836,
      "step": 11810
    },
    {
      "epoch": 3.2137030995106035,
      "grad_norm": 1.6875,
      "learning_rate": 1.889016676250719e-05,
      "loss": 0.0921,
      "step": 11820
    },
    {
      "epoch": 3.21642196846112,
      "grad_norm": 1.125,
      "learning_rate": 1.886141460609546e-05,
      "loss": 0.0823,
      "step": 11830
    },
    {
      "epoch": 3.219140837411637,
      "grad_norm": 2.1875,
      "learning_rate": 1.8832662449683726e-05,
      "loss": 0.1082,
      "step": 11840
    },
    {
      "epoch": 3.2218597063621535,
      "grad_norm": 0.578125,
      "learning_rate": 1.8803910293271994e-05,
      "loss": 0.0576,
      "step": 11850
    },
    {
      "epoch": 3.2218597063621535,
      "eval_loss": 0.11545312404632568,
      "eval_runtime": 100.9335,
      "eval_samples_per_second": 4.954,
      "eval_steps_per_second": 0.159,
      "step": 11850
    },
    {
      "epoch": 3.22457857531267,
      "grad_norm": 1.1640625,
      "learning_rate": 1.8775158136860264e-05,
      "loss": 0.0852,
      "step": 11860
    },
    {
      "epoch": 3.2272974442631863,
      "grad_norm": 0.6953125,
      "learning_rate": 1.8746405980448535e-05,
      "loss": 0.0832,
      "step": 11870
    },
    {
      "epoch": 3.230016313213703,
      "grad_norm": 2.328125,
      "learning_rate": 1.8717653824036806e-05,
      "loss": 0.1011,
      "step": 11880
    },
    {
      "epoch": 3.2327351821642196,
      "grad_norm": 1.75,
      "learning_rate": 1.8688901667625073e-05,
      "loss": 0.1066,
      "step": 11890
    },
    {
      "epoch": 3.2354540511147363,
      "grad_norm": 1.6484375,
      "learning_rate": 1.8660149511213344e-05,
      "loss": 0.0767,
      "step": 11900
    },
    {
      "epoch": 3.2354540511147363,
      "eval_loss": 0.11609374731779099,
      "eval_runtime": 106.9879,
      "eval_samples_per_second": 4.673,
      "eval_steps_per_second": 0.15,
      "step": 11900
    },
    {
      "epoch": 3.238172920065253,
      "grad_norm": 0.83984375,
      "learning_rate": 1.863139735480161e-05,
      "loss": 0.0727,
      "step": 11910
    },
    {
      "epoch": 3.2408917890157696,
      "grad_norm": 1.5703125,
      "learning_rate": 1.8602645198389882e-05,
      "loss": 0.1141,
      "step": 11920
    },
    {
      "epoch": 3.2436106579662862,
      "grad_norm": 3.09375,
      "learning_rate": 1.857389304197815e-05,
      "loss": 0.0951,
      "step": 11930
    },
    {
      "epoch": 3.2463295269168024,
      "grad_norm": 3.015625,
      "learning_rate": 1.8545140885566417e-05,
      "loss": 0.0895,
      "step": 11940
    },
    {
      "epoch": 3.249048395867319,
      "grad_norm": 1.171875,
      "learning_rate": 1.8516388729154687e-05,
      "loss": 0.0858,
      "step": 11950
    },
    {
      "epoch": 3.249048395867319,
      "eval_loss": 0.12160155922174454,
      "eval_runtime": 101.1044,
      "eval_samples_per_second": 4.945,
      "eval_steps_per_second": 0.158,
      "step": 11950
    },
    {
      "epoch": 3.2517672648178357,
      "grad_norm": 0.890625,
      "learning_rate": 1.8487636572742955e-05,
      "loss": 0.0752,
      "step": 11960
    },
    {
      "epoch": 3.2544861337683524,
      "grad_norm": 1.0234375,
      "learning_rate": 1.8458884416331225e-05,
      "loss": 0.0994,
      "step": 11970
    },
    {
      "epoch": 3.257205002718869,
      "grad_norm": 2.265625,
      "learning_rate": 1.8430132259919493e-05,
      "loss": 0.1012,
      "step": 11980
    },
    {
      "epoch": 3.2599238716693857,
      "grad_norm": 1.21875,
      "learning_rate": 1.8401380103507763e-05,
      "loss": 0.0898,
      "step": 11990
    },
    {
      "epoch": 3.262642740619902,
      "grad_norm": 1.6015625,
      "learning_rate": 1.8372627947096034e-05,
      "loss": 0.0948,
      "step": 12000
    },
    {
      "epoch": 3.262642740619902,
      "eval_loss": 0.1218281239271164,
      "eval_runtime": 100.7909,
      "eval_samples_per_second": 4.961,
      "eval_steps_per_second": 0.159,
      "step": 12000
    },
    {
      "epoch": 3.2653616095704185,
      "grad_norm": 1.78125,
      "learning_rate": 1.8343875790684305e-05,
      "loss": 0.0908,
      "step": 12010
    },
    {
      "epoch": 3.268080478520935,
      "grad_norm": 1.3515625,
      "learning_rate": 1.8315123634272572e-05,
      "loss": 0.0917,
      "step": 12020
    },
    {
      "epoch": 3.270799347471452,
      "grad_norm": 2.390625,
      "learning_rate": 1.828637147786084e-05,
      "loss": 0.0868,
      "step": 12030
    },
    {
      "epoch": 3.2735182164219685,
      "grad_norm": 1.1875,
      "learning_rate": 1.825761932144911e-05,
      "loss": 0.0797,
      "step": 12040
    },
    {
      "epoch": 3.276237085372485,
      "grad_norm": 1.6328125,
      "learning_rate": 1.8228867165037377e-05,
      "loss": 0.1085,
      "step": 12050
    },
    {
      "epoch": 3.276237085372485,
      "eval_loss": 0.11418750137090683,
      "eval_runtime": 100.4883,
      "eval_samples_per_second": 4.976,
      "eval_steps_per_second": 0.159,
      "step": 12050
    },
    {
      "epoch": 3.278955954323002,
      "grad_norm": 2.453125,
      "learning_rate": 1.8200115008625648e-05,
      "loss": 0.0951,
      "step": 12060
    },
    {
      "epoch": 3.2816748232735184,
      "grad_norm": 2.25,
      "learning_rate": 1.8171362852213915e-05,
      "loss": 0.0651,
      "step": 12070
    },
    {
      "epoch": 3.2843936922240347,
      "grad_norm": 1.3125,
      "learning_rate": 1.8142610695802186e-05,
      "loss": 0.1002,
      "step": 12080
    },
    {
      "epoch": 3.2871125611745513,
      "grad_norm": 1.1640625,
      "learning_rate": 1.8113858539390453e-05,
      "loss": 0.078,
      "step": 12090
    },
    {
      "epoch": 3.289831430125068,
      "grad_norm": 0.97265625,
      "learning_rate": 1.8085106382978724e-05,
      "loss": 0.1055,
      "step": 12100
    },
    {
      "epoch": 3.289831430125068,
      "eval_loss": 0.11823437362909317,
      "eval_runtime": 100.3493,
      "eval_samples_per_second": 4.983,
      "eval_steps_per_second": 0.159,
      "step": 12100
    },
    {
      "epoch": 3.2925502990755846,
      "grad_norm": 1.1953125,
      "learning_rate": 1.8056354226566995e-05,
      "loss": 0.0718,
      "step": 12110
    },
    {
      "epoch": 3.2952691680261013,
      "grad_norm": 0.35546875,
      "learning_rate": 1.8027602070155262e-05,
      "loss": 0.0735,
      "step": 12120
    },
    {
      "epoch": 3.297988036976618,
      "grad_norm": 2.3125,
      "learning_rate": 1.7998849913743533e-05,
      "loss": 0.0883,
      "step": 12130
    },
    {
      "epoch": 3.300706905927134,
      "grad_norm": 2.125,
      "learning_rate": 1.79700977573318e-05,
      "loss": 0.107,
      "step": 12140
    },
    {
      "epoch": 3.3034257748776508,
      "grad_norm": 0.69921875,
      "learning_rate": 1.794134560092007e-05,
      "loss": 0.0875,
      "step": 12150
    },
    {
      "epoch": 3.3034257748776508,
      "eval_loss": 0.11024218797683716,
      "eval_runtime": 100.1165,
      "eval_samples_per_second": 4.994,
      "eval_steps_per_second": 0.16,
      "step": 12150
    },
    {
      "epoch": 3.3061446438281674,
      "grad_norm": 1.75,
      "learning_rate": 1.7912593444508338e-05,
      "loss": 0.0753,
      "step": 12160
    },
    {
      "epoch": 3.308863512778684,
      "grad_norm": 1.90625,
      "learning_rate": 1.788384128809661e-05,
      "loss": 0.1104,
      "step": 12170
    },
    {
      "epoch": 3.3115823817292007,
      "grad_norm": 1.9765625,
      "learning_rate": 1.7855089131684876e-05,
      "loss": 0.0897,
      "step": 12180
    },
    {
      "epoch": 3.3143012506797174,
      "grad_norm": 2.109375,
      "learning_rate": 1.7826336975273147e-05,
      "loss": 0.0669,
      "step": 12190
    },
    {
      "epoch": 3.317020119630234,
      "grad_norm": 1.65625,
      "learning_rate": 1.7797584818861414e-05,
      "loss": 0.095,
      "step": 12200
    },
    {
      "epoch": 3.317020119630234,
      "eval_loss": 0.11778125166893005,
      "eval_runtime": 99.3519,
      "eval_samples_per_second": 5.033,
      "eval_steps_per_second": 0.161,
      "step": 12200
    },
    {
      "epoch": 3.3197389885807502,
      "grad_norm": 1.1328125,
      "learning_rate": 1.7768832662449685e-05,
      "loss": 0.0786,
      "step": 12210
    },
    {
      "epoch": 3.322457857531267,
      "grad_norm": 0.80859375,
      "learning_rate": 1.7740080506037952e-05,
      "loss": 0.0803,
      "step": 12220
    },
    {
      "epoch": 3.3251767264817835,
      "grad_norm": 0.67578125,
      "learning_rate": 1.7711328349626223e-05,
      "loss": 0.0879,
      "step": 12230
    },
    {
      "epoch": 3.3278955954323,
      "grad_norm": 1.265625,
      "learning_rate": 1.7682576193214494e-05,
      "loss": 0.0947,
      "step": 12240
    },
    {
      "epoch": 3.330614464382817,
      "grad_norm": 1.140625,
      "learning_rate": 1.765382403680276e-05,
      "loss": 0.0923,
      "step": 12250
    },
    {
      "epoch": 3.330614464382817,
      "eval_loss": 0.11961718648672104,
      "eval_runtime": 99.7187,
      "eval_samples_per_second": 5.014,
      "eval_steps_per_second": 0.16,
      "step": 12250
    },
    {
      "epoch": 3.3333333333333335,
      "grad_norm": 2.046875,
      "learning_rate": 1.762507188039103e-05,
      "loss": 0.099,
      "step": 12260
    },
    {
      "epoch": 3.3360522022838497,
      "grad_norm": 1.265625,
      "learning_rate": 1.75963197239793e-05,
      "loss": 0.0974,
      "step": 12270
    },
    {
      "epoch": 3.3387710712343663,
      "grad_norm": 2.34375,
      "learning_rate": 1.756756756756757e-05,
      "loss": 0.0821,
      "step": 12280
    },
    {
      "epoch": 3.341489940184883,
      "grad_norm": 0.8828125,
      "learning_rate": 1.7538815411155837e-05,
      "loss": 0.0711,
      "step": 12290
    },
    {
      "epoch": 3.3442088091353996,
      "grad_norm": 1.234375,
      "learning_rate": 1.7510063254744108e-05,
      "loss": 0.0885,
      "step": 12300
    },
    {
      "epoch": 3.3442088091353996,
      "eval_loss": 0.1148984357714653,
      "eval_runtime": 98.8695,
      "eval_samples_per_second": 5.057,
      "eval_steps_per_second": 0.162,
      "step": 12300
    },
    {
      "epoch": 3.3469276780859163,
      "grad_norm": 0.96875,
      "learning_rate": 1.7481311098332375e-05,
      "loss": 0.0789,
      "step": 12310
    },
    {
      "epoch": 3.349646547036433,
      "grad_norm": 2.5,
      "learning_rate": 1.7452558941920642e-05,
      "loss": 0.0914,
      "step": 12320
    },
    {
      "epoch": 3.3523654159869496,
      "grad_norm": 0.60546875,
      "learning_rate": 1.7423806785508913e-05,
      "loss": 0.0764,
      "step": 12330
    },
    {
      "epoch": 3.3550842849374662,
      "grad_norm": 1.3125,
      "learning_rate": 1.7395054629097184e-05,
      "loss": 0.0856,
      "step": 12340
    },
    {
      "epoch": 3.3578031538879825,
      "grad_norm": 1.9375,
      "learning_rate": 1.7366302472685454e-05,
      "loss": 0.1068,
      "step": 12350
    },
    {
      "epoch": 3.3578031538879825,
      "eval_loss": 0.11884374916553497,
      "eval_runtime": 98.0271,
      "eval_samples_per_second": 5.101,
      "eval_steps_per_second": 0.163,
      "step": 12350
    },
    {
      "epoch": 3.360522022838499,
      "grad_norm": 2.125,
      "learning_rate": 1.733755031627372e-05,
      "loss": 0.0854,
      "step": 12360
    },
    {
      "epoch": 3.3632408917890158,
      "grad_norm": 2.28125,
      "learning_rate": 1.7308798159861992e-05,
      "loss": 0.0902,
      "step": 12370
    },
    {
      "epoch": 3.3659597607395324,
      "grad_norm": 2.40625,
      "learning_rate": 1.728004600345026e-05,
      "loss": 0.0937,
      "step": 12380
    },
    {
      "epoch": 3.368678629690049,
      "grad_norm": 2.328125,
      "learning_rate": 1.725129384703853e-05,
      "loss": 0.0905,
      "step": 12390
    },
    {
      "epoch": 3.3713974986405657,
      "grad_norm": 1.375,
      "learning_rate": 1.7222541690626798e-05,
      "loss": 0.084,
      "step": 12400
    },
    {
      "epoch": 3.3713974986405657,
      "eval_loss": 0.1155703142285347,
      "eval_runtime": 114.368,
      "eval_samples_per_second": 4.372,
      "eval_steps_per_second": 0.14,
      "step": 12400
    },
    {
      "epoch": 3.374116367591082,
      "grad_norm": 1.84375,
      "learning_rate": 1.7193789534215065e-05,
      "loss": 0.0965,
      "step": 12410
    },
    {
      "epoch": 3.3768352365415986,
      "grad_norm": 0.93359375,
      "learning_rate": 1.7165037377803336e-05,
      "loss": 0.1033,
      "step": 12420
    },
    {
      "epoch": 3.379554105492115,
      "grad_norm": 1.3984375,
      "learning_rate": 1.7136285221391603e-05,
      "loss": 0.0917,
      "step": 12430
    },
    {
      "epoch": 3.382272974442632,
      "grad_norm": 0.9375,
      "learning_rate": 1.7107533064979874e-05,
      "loss": 0.0895,
      "step": 12440
    },
    {
      "epoch": 3.3849918433931485,
      "grad_norm": 1.0,
      "learning_rate": 1.707878090856814e-05,
      "loss": 0.0801,
      "step": 12450
    },
    {
      "epoch": 3.3849918433931485,
      "eval_loss": 0.1155078113079071,
      "eval_runtime": 105.1273,
      "eval_samples_per_second": 4.756,
      "eval_steps_per_second": 0.152,
      "step": 12450
    },
    {
      "epoch": 3.387710712343665,
      "grad_norm": 2.359375,
      "learning_rate": 1.7050028752156412e-05,
      "loss": 0.0873,
      "step": 12460
    },
    {
      "epoch": 3.390429581294182,
      "grad_norm": 2.140625,
      "learning_rate": 1.7021276595744682e-05,
      "loss": 0.0908,
      "step": 12470
    },
    {
      "epoch": 3.393148450244698,
      "grad_norm": 1.0703125,
      "learning_rate": 1.6992524439332953e-05,
      "loss": 0.0938,
      "step": 12480
    },
    {
      "epoch": 3.3958673191952147,
      "grad_norm": 2.671875,
      "learning_rate": 1.696377228292122e-05,
      "loss": 0.0935,
      "step": 12490
    },
    {
      "epoch": 3.3985861881457313,
      "grad_norm": 2.0625,
      "learning_rate": 1.6935020126509488e-05,
      "loss": 0.0961,
      "step": 12500
    },
    {
      "epoch": 3.3985861881457313,
      "eval_loss": 0.11375781148672104,
      "eval_runtime": 102.2884,
      "eval_samples_per_second": 4.888,
      "eval_steps_per_second": 0.156,
      "step": 12500
    },
    {
      "epoch": 3.401305057096248,
      "grad_norm": 0.984375,
      "learning_rate": 1.690626797009776e-05,
      "loss": 0.0777,
      "step": 12510
    },
    {
      "epoch": 3.4040239260467646,
      "grad_norm": 1.09375,
      "learning_rate": 1.6877515813686026e-05,
      "loss": 0.1043,
      "step": 12520
    },
    {
      "epoch": 3.4067427949972813,
      "grad_norm": 1.109375,
      "learning_rate": 1.6848763657274297e-05,
      "loss": 0.1202,
      "step": 12530
    },
    {
      "epoch": 3.4094616639477975,
      "grad_norm": 3.078125,
      "learning_rate": 1.6820011500862564e-05,
      "loss": 0.0727,
      "step": 12540
    },
    {
      "epoch": 3.412180532898314,
      "grad_norm": 1.640625,
      "learning_rate": 1.6791259344450835e-05,
      "loss": 0.0843,
      "step": 12550
    },
    {
      "epoch": 3.412180532898314,
      "eval_loss": 0.11437500268220901,
      "eval_runtime": 102.5876,
      "eval_samples_per_second": 4.874,
      "eval_steps_per_second": 0.156,
      "step": 12550
    },
    {
      "epoch": 3.414899401848831,
      "grad_norm": 0.73828125,
      "learning_rate": 1.6762507188039102e-05,
      "loss": 0.0921,
      "step": 12560
    },
    {
      "epoch": 3.4176182707993474,
      "grad_norm": 2.03125,
      "learning_rate": 1.6733755031627373e-05,
      "loss": 0.0691,
      "step": 12570
    },
    {
      "epoch": 3.420337139749864,
      "grad_norm": 1.640625,
      "learning_rate": 1.6705002875215643e-05,
      "loss": 0.0749,
      "step": 12580
    },
    {
      "epoch": 3.4230560087003807,
      "grad_norm": 1.71875,
      "learning_rate": 1.6676250718803914e-05,
      "loss": 0.1031,
      "step": 12590
    },
    {
      "epoch": 3.4257748776508974,
      "grad_norm": 1.390625,
      "learning_rate": 1.664749856239218e-05,
      "loss": 0.0853,
      "step": 12600
    },
    {
      "epoch": 3.4257748776508974,
      "eval_loss": 0.10890624672174454,
      "eval_runtime": 100.5489,
      "eval_samples_per_second": 4.973,
      "eval_steps_per_second": 0.159,
      "step": 12600
    },
    {
      "epoch": 3.428493746601414,
      "grad_norm": 2.5625,
      "learning_rate": 1.661874640598045e-05,
      "loss": 0.099,
      "step": 12610
    },
    {
      "epoch": 3.4312126155519302,
      "grad_norm": 0.7265625,
      "learning_rate": 1.658999424956872e-05,
      "loss": 0.0767,
      "step": 12620
    },
    {
      "epoch": 3.433931484502447,
      "grad_norm": 2.0,
      "learning_rate": 1.6561242093156987e-05,
      "loss": 0.08,
      "step": 12630
    },
    {
      "epoch": 3.4366503534529635,
      "grad_norm": 1.2265625,
      "learning_rate": 1.6532489936745257e-05,
      "loss": 0.1027,
      "step": 12640
    },
    {
      "epoch": 3.43936922240348,
      "grad_norm": 1.6640625,
      "learning_rate": 1.6503737780333525e-05,
      "loss": 0.0779,
      "step": 12650
    },
    {
      "epoch": 3.43936922240348,
      "eval_loss": 0.1148984357714653,
      "eval_runtime": 100.2888,
      "eval_samples_per_second": 4.986,
      "eval_steps_per_second": 0.16,
      "step": 12650
    },
    {
      "epoch": 3.442088091353997,
      "grad_norm": 1.6171875,
      "learning_rate": 1.6474985623921795e-05,
      "loss": 0.0707,
      "step": 12660
    },
    {
      "epoch": 3.4448069603045135,
      "grad_norm": 2.78125,
      "learning_rate": 1.6446233467510063e-05,
      "loss": 0.0876,
      "step": 12670
    },
    {
      "epoch": 3.4475258292550297,
      "grad_norm": 1.7421875,
      "learning_rate": 1.6417481311098333e-05,
      "loss": 0.0912,
      "step": 12680
    },
    {
      "epoch": 3.4502446982055464,
      "grad_norm": 1.09375,
      "learning_rate": 1.63887291546866e-05,
      "loss": 0.0866,
      "step": 12690
    },
    {
      "epoch": 3.452963567156063,
      "grad_norm": 1.578125,
      "learning_rate": 1.635997699827487e-05,
      "loss": 0.0964,
      "step": 12700
    },
    {
      "epoch": 3.452963567156063,
      "eval_loss": 0.12143749743700027,
      "eval_runtime": 98.6915,
      "eval_samples_per_second": 5.066,
      "eval_steps_per_second": 0.162,
      "step": 12700
    },
    {
      "epoch": 3.4556824361065797,
      "grad_norm": 1.875,
      "learning_rate": 1.6331224841863142e-05,
      "loss": 0.0997,
      "step": 12710
    },
    {
      "epoch": 3.4584013050570963,
      "grad_norm": 1.609375,
      "learning_rate": 1.630247268545141e-05,
      "loss": 0.0884,
      "step": 12720
    },
    {
      "epoch": 3.461120174007613,
      "grad_norm": 0.5390625,
      "learning_rate": 1.627372052903968e-05,
      "loss": 0.0698,
      "step": 12730
    },
    {
      "epoch": 3.4638390429581296,
      "grad_norm": 1.4921875,
      "learning_rate": 1.6244968372627947e-05,
      "loss": 0.0773,
      "step": 12740
    },
    {
      "epoch": 3.466557911908646,
      "grad_norm": 0.83984375,
      "learning_rate": 1.6216216216216218e-05,
      "loss": 0.07,
      "step": 12750
    },
    {
      "epoch": 3.466557911908646,
      "eval_loss": 0.11067187786102295,
      "eval_runtime": 100.3663,
      "eval_samples_per_second": 4.982,
      "eval_steps_per_second": 0.159,
      "step": 12750
    },
    {
      "epoch": 3.4692767808591625,
      "grad_norm": 1.96875,
      "learning_rate": 1.6187464059804485e-05,
      "loss": 0.0695,
      "step": 12760
    },
    {
      "epoch": 3.471995649809679,
      "grad_norm": 2.234375,
      "learning_rate": 1.6158711903392756e-05,
      "loss": 0.1021,
      "step": 12770
    },
    {
      "epoch": 3.4747145187601958,
      "grad_norm": 1.53125,
      "learning_rate": 1.6129959746981023e-05,
      "loss": 0.0872,
      "step": 12780
    },
    {
      "epoch": 3.4774333877107124,
      "grad_norm": 1.3984375,
      "learning_rate": 1.610120759056929e-05,
      "loss": 0.0825,
      "step": 12790
    },
    {
      "epoch": 3.480152256661229,
      "grad_norm": 0.82421875,
      "learning_rate": 1.607245543415756e-05,
      "loss": 0.0664,
      "step": 12800
    },
    {
      "epoch": 3.480152256661229,
      "eval_loss": 0.12171874940395355,
      "eval_runtime": 98.5246,
      "eval_samples_per_second": 5.075,
      "eval_steps_per_second": 0.162,
      "step": 12800
    },
    {
      "epoch": 3.4828711256117453,
      "grad_norm": 1.4296875,
      "learning_rate": 1.6043703277745832e-05,
      "loss": 0.0666,
      "step": 12810
    },
    {
      "epoch": 3.485589994562262,
      "grad_norm": 1.2890625,
      "learning_rate": 1.6014951121334103e-05,
      "loss": 0.0911,
      "step": 12820
    },
    {
      "epoch": 3.4883088635127786,
      "grad_norm": 2.03125,
      "learning_rate": 1.598619896492237e-05,
      "loss": 0.1218,
      "step": 12830
    },
    {
      "epoch": 3.4910277324632952,
      "grad_norm": 2.40625,
      "learning_rate": 1.595744680851064e-05,
      "loss": 0.1136,
      "step": 12840
    },
    {
      "epoch": 3.493746601413812,
      "grad_norm": 1.953125,
      "learning_rate": 1.5928694652098908e-05,
      "loss": 0.0789,
      "step": 12850
    },
    {
      "epoch": 3.493746601413812,
      "eval_loss": 0.12339062243700027,
      "eval_runtime": 98.0292,
      "eval_samples_per_second": 5.101,
      "eval_steps_per_second": 0.163,
      "step": 12850
    },
    {
      "epoch": 3.4964654703643285,
      "grad_norm": 1.484375,
      "learning_rate": 1.589994249568718e-05,
      "loss": 0.067,
      "step": 12860
    },
    {
      "epoch": 3.499184339314845,
      "grad_norm": 1.3671875,
      "learning_rate": 1.5871190339275446e-05,
      "loss": 0.1004,
      "step": 12870
    },
    {
      "epoch": 3.501903208265362,
      "grad_norm": 1.828125,
      "learning_rate": 1.5842438182863714e-05,
      "loss": 0.0856,
      "step": 12880
    },
    {
      "epoch": 3.504622077215878,
      "grad_norm": 1.0703125,
      "learning_rate": 1.5813686026451984e-05,
      "loss": 0.0887,
      "step": 12890
    },
    {
      "epoch": 3.5073409461663947,
      "grad_norm": 3.390625,
      "learning_rate": 1.578493387004025e-05,
      "loss": 0.089,
      "step": 12900
    },
    {
      "epoch": 3.5073409461663947,
      "eval_loss": 0.11016406118869781,
      "eval_runtime": 104.7145,
      "eval_samples_per_second": 4.775,
      "eval_steps_per_second": 0.153,
      "step": 12900
    },
    {
      "epoch": 3.5100598151169113,
      "grad_norm": 0.953125,
      "learning_rate": 1.5756181713628522e-05,
      "loss": 0.0942,
      "step": 12910
    },
    {
      "epoch": 3.512778684067428,
      "grad_norm": 0.96875,
      "learning_rate": 1.572742955721679e-05,
      "loss": 0.0835,
      "step": 12920
    },
    {
      "epoch": 3.5154975530179446,
      "grad_norm": 0.90625,
      "learning_rate": 1.569867740080506e-05,
      "loss": 0.0826,
      "step": 12930
    },
    {
      "epoch": 3.518216421968461,
      "grad_norm": 1.328125,
      "learning_rate": 1.566992524439333e-05,
      "loss": 0.0828,
      "step": 12940
    },
    {
      "epoch": 3.5209352909189775,
      "grad_norm": 2.78125,
      "learning_rate": 1.56411730879816e-05,
      "loss": 0.0794,
      "step": 12950
    },
    {
      "epoch": 3.5209352909189775,
      "eval_loss": 0.12160155922174454,
      "eval_runtime": 98.375,
      "eval_samples_per_second": 5.083,
      "eval_steps_per_second": 0.163,
      "step": 12950
    },
    {
      "epoch": 3.523654159869494,
      "grad_norm": 1.5546875,
      "learning_rate": 1.561242093156987e-05,
      "loss": 0.0945,
      "step": 12960
    },
    {
      "epoch": 3.526373028820011,
      "grad_norm": 1.046875,
      "learning_rate": 1.558366877515814e-05,
      "loss": 0.0954,
      "step": 12970
    },
    {
      "epoch": 3.5290918977705275,
      "grad_norm": 0.6796875,
      "learning_rate": 1.5554916618746407e-05,
      "loss": 0.0944,
      "step": 12980
    },
    {
      "epoch": 3.531810766721044,
      "grad_norm": 1.265625,
      "learning_rate": 1.5526164462334674e-05,
      "loss": 0.1009,
      "step": 12990
    },
    {
      "epoch": 3.5345296356715608,
      "grad_norm": 1.3984375,
      "learning_rate": 1.5497412305922945e-05,
      "loss": 0.0872,
      "step": 13000
    },
    {
      "epoch": 3.5345296356715608,
      "eval_loss": 0.11217968910932541,
      "eval_runtime": 98.9806,
      "eval_samples_per_second": 5.051,
      "eval_steps_per_second": 0.162,
      "step": 13000
    },
    {
      "epoch": 3.5372485046220774,
      "grad_norm": 1.984375,
      "learning_rate": 1.5468660149511212e-05,
      "loss": 0.0924,
      "step": 13010
    },
    {
      "epoch": 3.539967373572594,
      "grad_norm": 0.95703125,
      "learning_rate": 1.5439907993099483e-05,
      "loss": 0.0879,
      "step": 13020
    },
    {
      "epoch": 3.5426862425231103,
      "grad_norm": 1.296875,
      "learning_rate": 1.541115583668775e-05,
      "loss": 0.0774,
      "step": 13030
    },
    {
      "epoch": 3.545405111473627,
      "grad_norm": 1.0859375,
      "learning_rate": 1.538240368027602e-05,
      "loss": 0.0862,
      "step": 13040
    },
    {
      "epoch": 3.5481239804241436,
      "grad_norm": 1.1015625,
      "learning_rate": 1.5353651523864292e-05,
      "loss": 0.0718,
      "step": 13050
    },
    {
      "epoch": 3.5481239804241436,
      "eval_loss": 0.11139843612909317,
      "eval_runtime": 98.3108,
      "eval_samples_per_second": 5.086,
      "eval_steps_per_second": 0.163,
      "step": 13050
    },
    {
      "epoch": 3.55084284937466,
      "grad_norm": 1.3359375,
      "learning_rate": 1.5324899367452562e-05,
      "loss": 0.0885,
      "step": 13060
    },
    {
      "epoch": 3.553561718325177,
      "grad_norm": 2.40625,
      "learning_rate": 1.529614721104083e-05,
      "loss": 0.0917,
      "step": 13070
    },
    {
      "epoch": 3.556280587275693,
      "grad_norm": 1.9453125,
      "learning_rate": 1.5267395054629097e-05,
      "loss": 0.09,
      "step": 13080
    },
    {
      "epoch": 3.5589994562262097,
      "grad_norm": 1.3515625,
      "learning_rate": 1.5238642898217368e-05,
      "loss": 0.0926,
      "step": 13090
    },
    {
      "epoch": 3.5617183251767264,
      "grad_norm": 1.9453125,
      "learning_rate": 1.5209890741805635e-05,
      "loss": 0.1001,
      "step": 13100
    },
    {
      "epoch": 3.5617183251767264,
      "eval_loss": 0.11934375017881393,
      "eval_runtime": 97.2497,
      "eval_samples_per_second": 5.141,
      "eval_steps_per_second": 0.165,
      "step": 13100
    },
    {
      "epoch": 3.564437194127243,
      "grad_norm": 2.359375,
      "learning_rate": 1.5181138585393906e-05,
      "loss": 0.0924,
      "step": 13110
    },
    {
      "epoch": 3.5671560630777597,
      "grad_norm": 1.984375,
      "learning_rate": 1.5152386428982173e-05,
      "loss": 0.0933,
      "step": 13120
    },
    {
      "epoch": 3.5698749320282763,
      "grad_norm": 1.5703125,
      "learning_rate": 1.5123634272570444e-05,
      "loss": 0.0789,
      "step": 13130
    },
    {
      "epoch": 3.572593800978793,
      "grad_norm": 2.5,
      "learning_rate": 1.5094882116158713e-05,
      "loss": 0.0945,
      "step": 13140
    },
    {
      "epoch": 3.5753126699293096,
      "grad_norm": 3.15625,
      "learning_rate": 1.5066129959746984e-05,
      "loss": 0.1064,
      "step": 13150
    },
    {
      "epoch": 3.5753126699293096,
      "eval_loss": 0.11125781387090683,
      "eval_runtime": 97.8357,
      "eval_samples_per_second": 5.111,
      "eval_steps_per_second": 0.164,
      "step": 13150
    },
    {
      "epoch": 3.578031538879826,
      "grad_norm": 1.3046875,
      "learning_rate": 1.503737780333525e-05,
      "loss": 0.1032,
      "step": 13160
    },
    {
      "epoch": 3.5807504078303425,
      "grad_norm": 1.796875,
      "learning_rate": 1.5008625646923518e-05,
      "loss": 0.0818,
      "step": 13170
    },
    {
      "epoch": 3.583469276780859,
      "grad_norm": 0.91796875,
      "learning_rate": 1.4979873490511789e-05,
      "loss": 0.0907,
      "step": 13180
    },
    {
      "epoch": 3.586188145731376,
      "grad_norm": 3.453125,
      "learning_rate": 1.4951121334100058e-05,
      "loss": 0.1118,
      "step": 13190
    },
    {
      "epoch": 3.5889070146818924,
      "grad_norm": 1.921875,
      "learning_rate": 1.4922369177688329e-05,
      "loss": 0.0876,
      "step": 13200
    },
    {
      "epoch": 3.5889070146818924,
      "eval_loss": 0.11109375208616257,
      "eval_runtime": 96.2803,
      "eval_samples_per_second": 5.193,
      "eval_steps_per_second": 0.166,
      "step": 13200
    },
    {
      "epoch": 3.5916258836324086,
      "grad_norm": 1.109375,
      "learning_rate": 1.4893617021276596e-05,
      "loss": 0.1046,
      "step": 13210
    },
    {
      "epoch": 3.5943447525829253,
      "grad_norm": 0.85546875,
      "learning_rate": 1.4864864864864867e-05,
      "loss": 0.0732,
      "step": 13220
    },
    {
      "epoch": 3.597063621533442,
      "grad_norm": 0.83984375,
      "learning_rate": 1.4836112708453134e-05,
      "loss": 0.0898,
      "step": 13230
    },
    {
      "epoch": 3.5997824904839586,
      "grad_norm": 1.234375,
      "learning_rate": 1.4807360552041405e-05,
      "loss": 0.0912,
      "step": 13240
    },
    {
      "epoch": 3.6025013594344752,
      "grad_norm": 1.7109375,
      "learning_rate": 1.4778608395629674e-05,
      "loss": 0.0915,
      "step": 13250
    },
    {
      "epoch": 3.6025013594344752,
      "eval_loss": 0.11797656118869781,
      "eval_runtime": 99.4952,
      "eval_samples_per_second": 5.025,
      "eval_steps_per_second": 0.161,
      "step": 13250
    },
    {
      "epoch": 3.605220228384992,
      "grad_norm": 2.359375,
      "learning_rate": 1.4749856239217941e-05,
      "loss": 0.1119,
      "step": 13260
    },
    {
      "epoch": 3.6079390973355085,
      "grad_norm": 1.2578125,
      "learning_rate": 1.4721104082806212e-05,
      "loss": 0.0956,
      "step": 13270
    },
    {
      "epoch": 3.610657966286025,
      "grad_norm": 1.2578125,
      "learning_rate": 1.4692351926394479e-05,
      "loss": 0.0938,
      "step": 13280
    },
    {
      "epoch": 3.613376835236542,
      "grad_norm": 1.2890625,
      "learning_rate": 1.466359976998275e-05,
      "loss": 0.0962,
      "step": 13290
    },
    {
      "epoch": 3.616095704187058,
      "grad_norm": 2.0625,
      "learning_rate": 1.4634847613571017e-05,
      "loss": 0.0967,
      "step": 13300
    },
    {
      "epoch": 3.616095704187058,
      "eval_loss": 0.1182265654206276,
      "eval_runtime": 99.7579,
      "eval_samples_per_second": 5.012,
      "eval_steps_per_second": 0.16,
      "step": 13300
    },
    {
      "epoch": 3.6188145731375747,
      "grad_norm": 0.91015625,
      "learning_rate": 1.4606095457159288e-05,
      "loss": 0.0929,
      "step": 13310
    },
    {
      "epoch": 3.6215334420880914,
      "grad_norm": 1.8046875,
      "learning_rate": 1.4577343300747557e-05,
      "loss": 0.1182,
      "step": 13320
    },
    {
      "epoch": 3.624252311038608,
      "grad_norm": 1.4609375,
      "learning_rate": 1.4548591144335827e-05,
      "loss": 0.0902,
      "step": 13330
    },
    {
      "epoch": 3.6269711799891247,
      "grad_norm": 2.625,
      "learning_rate": 1.4519838987924095e-05,
      "loss": 0.0797,
      "step": 13340
    },
    {
      "epoch": 3.629690048939641,
      "grad_norm": 1.453125,
      "learning_rate": 1.4491086831512365e-05,
      "loss": 0.0818,
      "step": 13350
    },
    {
      "epoch": 3.629690048939641,
      "eval_loss": 0.11707030981779099,
      "eval_runtime": 98.347,
      "eval_samples_per_second": 5.084,
      "eval_steps_per_second": 0.163,
      "step": 13350
    },
    {
      "epoch": 3.6324089178901575,
      "grad_norm": 1.0546875,
      "learning_rate": 1.4462334675100633e-05,
      "loss": 0.0775,
      "step": 13360
    },
    {
      "epoch": 3.635127786840674,
      "grad_norm": 1.078125,
      "learning_rate": 1.4433582518688902e-05,
      "loss": 0.0862,
      "step": 13370
    },
    {
      "epoch": 3.637846655791191,
      "grad_norm": 1.8046875,
      "learning_rate": 1.4404830362277172e-05,
      "loss": 0.0865,
      "step": 13380
    },
    {
      "epoch": 3.6405655247417075,
      "grad_norm": 0.796875,
      "learning_rate": 1.437607820586544e-05,
      "loss": 0.1073,
      "step": 13390
    },
    {
      "epoch": 3.643284393692224,
      "grad_norm": 0.79296875,
      "learning_rate": 1.434732604945371e-05,
      "loss": 0.1043,
      "step": 13400
    },
    {
      "epoch": 3.643284393692224,
      "eval_loss": 0.1135859340429306,
      "eval_runtime": 114.6564,
      "eval_samples_per_second": 4.361,
      "eval_steps_per_second": 0.14,
      "step": 13400
    },
    {
      "epoch": 3.6460032626427408,
      "grad_norm": 0.91796875,
      "learning_rate": 1.4318573893041978e-05,
      "loss": 0.0988,
      "step": 13410
    },
    {
      "epoch": 3.6487221315932574,
      "grad_norm": 1.71875,
      "learning_rate": 1.4289821736630248e-05,
      "loss": 0.0829,
      "step": 13420
    },
    {
      "epoch": 3.6514410005437736,
      "grad_norm": 3.0,
      "learning_rate": 1.4261069580218517e-05,
      "loss": 0.1206,
      "step": 13430
    },
    {
      "epoch": 3.6541598694942903,
      "grad_norm": 1.4765625,
      "learning_rate": 1.4232317423806788e-05,
      "loss": 0.088,
      "step": 13440
    },
    {
      "epoch": 3.656878738444807,
      "grad_norm": 0.95703125,
      "learning_rate": 1.4203565267395055e-05,
      "loss": 0.1064,
      "step": 13450
    },
    {
      "epoch": 3.656878738444807,
      "eval_loss": 0.11445312201976776,
      "eval_runtime": 103.1864,
      "eval_samples_per_second": 4.846,
      "eval_steps_per_second": 0.155,
      "step": 13450
    },
    {
      "epoch": 3.6595976073953236,
      "grad_norm": 1.1015625,
      "learning_rate": 1.4174813110983323e-05,
      "loss": 0.0839,
      "step": 13460
    },
    {
      "epoch": 3.6623164763458402,
      "grad_norm": 1.15625,
      "learning_rate": 1.4146060954571593e-05,
      "loss": 0.0978,
      "step": 13470
    },
    {
      "epoch": 3.6650353452963564,
      "grad_norm": 2.03125,
      "learning_rate": 1.4117308798159863e-05,
      "loss": 0.0994,
      "step": 13480
    },
    {
      "epoch": 3.667754214246873,
      "grad_norm": 2.1875,
      "learning_rate": 1.4088556641748133e-05,
      "loss": 0.0767,
      "step": 13490
    },
    {
      "epoch": 3.6704730831973897,
      "grad_norm": 1.265625,
      "learning_rate": 1.40598044853364e-05,
      "loss": 0.0734,
      "step": 13500
    },
    {
      "epoch": 3.6704730831973897,
      "eval_loss": 0.11408593505620956,
      "eval_runtime": 101.6027,
      "eval_samples_per_second": 4.921,
      "eval_steps_per_second": 0.157,
      "step": 13500
    },
    {
      "epoch": 3.6731919521479064,
      "grad_norm": 1.328125,
      "learning_rate": 1.4031052328924671e-05,
      "loss": 0.0813,
      "step": 13510
    },
    {
      "epoch": 3.675910821098423,
      "grad_norm": 1.9140625,
      "learning_rate": 1.4002300172512939e-05,
      "loss": 0.1014,
      "step": 13520
    },
    {
      "epoch": 3.6786296900489397,
      "grad_norm": 1.421875,
      "learning_rate": 1.397354801610121e-05,
      "loss": 0.0949,
      "step": 13530
    },
    {
      "epoch": 3.6813485589994563,
      "grad_norm": 0.67578125,
      "learning_rate": 1.3944795859689477e-05,
      "loss": 0.0866,
      "step": 13540
    },
    {
      "epoch": 3.684067427949973,
      "grad_norm": 1.0078125,
      "learning_rate": 1.3916043703277746e-05,
      "loss": 0.1056,
      "step": 13550
    },
    {
      "epoch": 3.684067427949973,
      "eval_loss": 0.10814844071865082,
      "eval_runtime": 102.1443,
      "eval_samples_per_second": 4.895,
      "eval_steps_per_second": 0.157,
      "step": 13550
    },
    {
      "epoch": 3.6867862969004896,
      "grad_norm": 0.494140625,
      "learning_rate": 1.3887291546866016e-05,
      "loss": 0.0877,
      "step": 13560
    },
    {
      "epoch": 3.689505165851006,
      "grad_norm": 1.1171875,
      "learning_rate": 1.3858539390454284e-05,
      "loss": 0.0945,
      "step": 13570
    },
    {
      "epoch": 3.6922240348015225,
      "grad_norm": 2.875,
      "learning_rate": 1.3829787234042554e-05,
      "loss": 0.1024,
      "step": 13580
    },
    {
      "epoch": 3.694942903752039,
      "grad_norm": 2.203125,
      "learning_rate": 1.3801035077630822e-05,
      "loss": 0.0927,
      "step": 13590
    },
    {
      "epoch": 3.697661772702556,
      "grad_norm": 1.8515625,
      "learning_rate": 1.3772282921219092e-05,
      "loss": 0.0664,
      "step": 13600
    },
    {
      "epoch": 3.697661772702556,
      "eval_loss": 0.10751562565565109,
      "eval_runtime": 100.6351,
      "eval_samples_per_second": 4.968,
      "eval_steps_per_second": 0.159,
      "step": 13600
    },
    {
      "epoch": 3.7003806416530725,
      "grad_norm": 1.6796875,
      "learning_rate": 1.3743530764807361e-05,
      "loss": 0.0911,
      "step": 13610
    },
    {
      "epoch": 3.7030995106035887,
      "grad_norm": 1.7890625,
      "learning_rate": 1.3714778608395632e-05,
      "loss": 0.0973,
      "step": 13620
    },
    {
      "epoch": 3.7058183795541053,
      "grad_norm": 1.421875,
      "learning_rate": 1.36860264519839e-05,
      "loss": 0.0796,
      "step": 13630
    },
    {
      "epoch": 3.708537248504622,
      "grad_norm": 1.5,
      "learning_rate": 1.3657274295572167e-05,
      "loss": 0.0781,
      "step": 13640
    },
    {
      "epoch": 3.7112561174551386,
      "grad_norm": 1.5859375,
      "learning_rate": 1.3628522139160437e-05,
      "loss": 0.1028,
      "step": 13650
    },
    {
      "epoch": 3.7112561174551386,
      "eval_loss": 0.11262500286102295,
      "eval_runtime": 100.3607,
      "eval_samples_per_second": 4.982,
      "eval_steps_per_second": 0.159,
      "step": 13650
    },
    {
      "epoch": 3.7139749864056553,
      "grad_norm": 1.4296875,
      "learning_rate": 1.3599769982748706e-05,
      "loss": 0.0782,
      "step": 13660
    },
    {
      "epoch": 3.716693855356172,
      "grad_norm": 1.0390625,
      "learning_rate": 1.3571017826336977e-05,
      "loss": 0.0912,
      "step": 13670
    },
    {
      "epoch": 3.7194127243066886,
      "grad_norm": 1.2734375,
      "learning_rate": 1.3542265669925244e-05,
      "loss": 0.0824,
      "step": 13680
    },
    {
      "epoch": 3.722131593257205,
      "grad_norm": 1.3203125,
      "learning_rate": 1.3513513513513515e-05,
      "loss": 0.0901,
      "step": 13690
    },
    {
      "epoch": 3.7248504622077214,
      "grad_norm": 1.671875,
      "learning_rate": 1.3484761357101782e-05,
      "loss": 0.0791,
      "step": 13700
    },
    {
      "epoch": 3.7248504622077214,
      "eval_loss": 0.1146249994635582,
      "eval_runtime": 115.3069,
      "eval_samples_per_second": 4.336,
      "eval_steps_per_second": 0.139,
      "step": 13700
    },
    {
      "epoch": 3.727569331158238,
      "grad_norm": 1.4140625,
      "learning_rate": 1.3456009200690053e-05,
      "loss": 0.0901,
      "step": 13710
    },
    {
      "epoch": 3.7302882001087547,
      "grad_norm": 1.3671875,
      "learning_rate": 1.3427257044278322e-05,
      "loss": 0.0847,
      "step": 13720
    },
    {
      "epoch": 3.7330070690592714,
      "grad_norm": 0.60546875,
      "learning_rate": 1.3398504887866593e-05,
      "loss": 0.0737,
      "step": 13730
    },
    {
      "epoch": 3.735725938009788,
      "grad_norm": 1.390625,
      "learning_rate": 1.336975273145486e-05,
      "loss": 0.0936,
      "step": 13740
    },
    {
      "epoch": 3.7384448069603047,
      "grad_norm": 1.328125,
      "learning_rate": 1.3341000575043127e-05,
      "loss": 0.0894,
      "step": 13750
    },
    {
      "epoch": 3.7384448069603047,
      "eval_loss": 0.11846093833446503,
      "eval_runtime": 98.1784,
      "eval_samples_per_second": 5.093,
      "eval_steps_per_second": 0.163,
      "step": 13750
    },
    {
      "epoch": 3.741163675910821,
      "grad_norm": 1.8671875,
      "learning_rate": 1.3312248418631398e-05,
      "loss": 0.1006,
      "step": 13760
    },
    {
      "epoch": 3.7438825448613375,
      "grad_norm": 0.98046875,
      "learning_rate": 1.3283496262219665e-05,
      "loss": 0.0844,
      "step": 13770
    },
    {
      "epoch": 3.746601413811854,
      "grad_norm": 0.93359375,
      "learning_rate": 1.3254744105807936e-05,
      "loss": 0.0905,
      "step": 13780
    },
    {
      "epoch": 3.749320282762371,
      "grad_norm": 1.53125,
      "learning_rate": 1.3225991949396205e-05,
      "loss": 0.0653,
      "step": 13790
    },
    {
      "epoch": 3.7520391517128875,
      "grad_norm": 2.0,
      "learning_rate": 1.3197239792984476e-05,
      "loss": 0.0948,
      "step": 13800
    },
    {
      "epoch": 3.7520391517128875,
      "eval_loss": 0.10938281565904617,
      "eval_runtime": 114.5714,
      "eval_samples_per_second": 4.364,
      "eval_steps_per_second": 0.14,
      "step": 13800
    },
    {
      "epoch": 3.754758020663404,
      "grad_norm": 1.84375,
      "learning_rate": 1.3168487636572743e-05,
      "loss": 0.0942,
      "step": 13810
    },
    {
      "epoch": 3.757476889613921,
      "grad_norm": 3.25,
      "learning_rate": 1.3139735480161014e-05,
      "loss": 0.0995,
      "step": 13820
    },
    {
      "epoch": 3.7601957585644374,
      "grad_norm": 1.203125,
      "learning_rate": 1.3110983323749281e-05,
      "loss": 0.0758,
      "step": 13830
    },
    {
      "epoch": 3.7629146275149536,
      "grad_norm": 1.6015625,
      "learning_rate": 1.308223116733755e-05,
      "loss": 0.0781,
      "step": 13840
    },
    {
      "epoch": 3.7656334964654703,
      "grad_norm": 2.0,
      "learning_rate": 1.3053479010925821e-05,
      "loss": 0.0823,
      "step": 13850
    },
    {
      "epoch": 3.7656334964654703,
      "eval_loss": 0.11556249856948853,
      "eval_runtime": 210.4666,
      "eval_samples_per_second": 2.376,
      "eval_steps_per_second": 0.076,
      "step": 13850
    },
    {
      "epoch": 3.768352365415987,
      "grad_norm": 1.6640625,
      "learning_rate": 1.3024726854514088e-05,
      "loss": 0.0791,
      "step": 13860
    },
    {
      "epoch": 3.7710712343665036,
      "grad_norm": 2.53125,
      "learning_rate": 1.2995974698102359e-05,
      "loss": 0.1054,
      "step": 13870
    },
    {
      "epoch": 3.7737901033170203,
      "grad_norm": 1.5546875,
      "learning_rate": 1.2967222541690626e-05,
      "loss": 0.1049,
      "step": 13880
    },
    {
      "epoch": 3.7765089722675365,
      "grad_norm": 1.1796875,
      "learning_rate": 1.2938470385278897e-05,
      "loss": 0.0646,
      "step": 13890
    },
    {
      "epoch": 3.779227841218053,
      "grad_norm": 4.09375,
      "learning_rate": 1.2909718228867166e-05,
      "loss": 0.0912,
      "step": 13900
    },
    {
      "epoch": 3.779227841218053,
      "eval_loss": 0.1266171932220459,
      "eval_runtime": 137.2403,
      "eval_samples_per_second": 3.643,
      "eval_steps_per_second": 0.117,
      "step": 13900
    },
    {
      "epoch": 3.7819467101685698,
      "grad_norm": 2.734375,
      "learning_rate": 1.2880966072455437e-05,
      "loss": 0.0784,
      "step": 13910
    },
    {
      "epoch": 3.7846655791190864,
      "grad_norm": 2.046875,
      "learning_rate": 1.2852213916043704e-05,
      "loss": 0.0913,
      "step": 13920
    },
    {
      "epoch": 3.787384448069603,
      "grad_norm": 2.78125,
      "learning_rate": 1.2823461759631971e-05,
      "loss": 0.0813,
      "step": 13930
    },
    {
      "epoch": 3.7901033170201197,
      "grad_norm": 1.0625,
      "learning_rate": 1.2794709603220242e-05,
      "loss": 0.1048,
      "step": 13940
    },
    {
      "epoch": 3.7928221859706364,
      "grad_norm": 0.95703125,
      "learning_rate": 1.2765957446808511e-05,
      "loss": 0.0713,
      "step": 13950
    },
    {
      "epoch": 3.7928221859706364,
      "eval_loss": 0.11633593589067459,
      "eval_runtime": 99.667,
      "eval_samples_per_second": 5.017,
      "eval_steps_per_second": 0.161,
      "step": 13950
    },
    {
      "epoch": 3.795541054921153,
      "grad_norm": 3.234375,
      "learning_rate": 1.2737205290396782e-05,
      "loss": 0.0995,
      "step": 13960
    },
    {
      "epoch": 3.798259923871669,
      "grad_norm": 1.1953125,
      "learning_rate": 1.2708453133985049e-05,
      "loss": 0.0881,
      "step": 13970
    },
    {
      "epoch": 3.800978792822186,
      "grad_norm": 3.484375,
      "learning_rate": 1.267970097757332e-05,
      "loss": 0.1248,
      "step": 13980
    },
    {
      "epoch": 3.8036976617727025,
      "grad_norm": 1.84375,
      "learning_rate": 1.2650948821161587e-05,
      "loss": 0.0961,
      "step": 13990
    },
    {
      "epoch": 3.806416530723219,
      "grad_norm": 0.55078125,
      "learning_rate": 1.2622196664749858e-05,
      "loss": 0.0745,
      "step": 14000
    },
    {
      "epoch": 3.806416530723219,
      "eval_loss": 0.11298437416553497,
      "eval_runtime": 118.6378,
      "eval_samples_per_second": 4.215,
      "eval_steps_per_second": 0.135,
      "step": 14000
    },
    {
      "epoch": 3.809135399673736,
      "grad_norm": 1.9921875,
      "learning_rate": 1.2593444508338125e-05,
      "loss": 0.0943,
      "step": 14010
    },
    {
      "epoch": 3.8118542686242525,
      "grad_norm": 2.203125,
      "learning_rate": 1.2564692351926394e-05,
      "loss": 0.1093,
      "step": 14020
    },
    {
      "epoch": 3.8145731375747687,
      "grad_norm": 1.8515625,
      "learning_rate": 1.2535940195514665e-05,
      "loss": 0.0891,
      "step": 14030
    },
    {
      "epoch": 3.8172920065252853,
      "grad_norm": 1.5,
      "learning_rate": 1.2507188039102932e-05,
      "loss": 0.0907,
      "step": 14040
    },
    {
      "epoch": 3.820010875475802,
      "grad_norm": 1.0390625,
      "learning_rate": 1.2478435882691203e-05,
      "loss": 0.0954,
      "step": 14050
    },
    {
      "epoch": 3.820010875475802,
      "eval_loss": 0.11301562190055847,
      "eval_runtime": 100.8117,
      "eval_samples_per_second": 4.96,
      "eval_steps_per_second": 0.159,
      "step": 14050
    },
    {
      "epoch": 3.8227297444263186,
      "grad_norm": 0.73046875,
      "learning_rate": 1.2449683726279472e-05,
      "loss": 0.0854,
      "step": 14060
    },
    {
      "epoch": 3.8254486133768353,
      "grad_norm": 2.59375,
      "learning_rate": 1.242093156986774e-05,
      "loss": 0.094,
      "step": 14070
    },
    {
      "epoch": 3.828167482327352,
      "grad_norm": 2.1875,
      "learning_rate": 1.239217941345601e-05,
      "loss": 0.091,
      "step": 14080
    },
    {
      "epoch": 3.8308863512778686,
      "grad_norm": 0.765625,
      "learning_rate": 1.2363427257044279e-05,
      "loss": 0.0798,
      "step": 14090
    },
    {
      "epoch": 3.8336052202283852,
      "grad_norm": 1.7578125,
      "learning_rate": 1.2334675100632548e-05,
      "loss": 0.0651,
      "step": 14100
    },
    {
      "epoch": 3.8336052202283852,
      "eval_loss": 0.11603125184774399,
      "eval_runtime": 99.6654,
      "eval_samples_per_second": 5.017,
      "eval_steps_per_second": 0.161,
      "step": 14100
    },
    {
      "epoch": 3.8363240891789014,
      "grad_norm": 0.5625,
      "learning_rate": 1.2305922944220817e-05,
      "loss": 0.0614,
      "step": 14110
    },
    {
      "epoch": 3.839042958129418,
      "grad_norm": 1.3359375,
      "learning_rate": 1.2277170787809086e-05,
      "loss": 0.0945,
      "step": 14120
    },
    {
      "epoch": 3.8417618270799347,
      "grad_norm": 2.34375,
      "learning_rate": 1.2248418631397355e-05,
      "loss": 0.0959,
      "step": 14130
    },
    {
      "epoch": 3.8444806960304514,
      "grad_norm": 3.8125,
      "learning_rate": 1.2219666474985626e-05,
      "loss": 0.0945,
      "step": 14140
    },
    {
      "epoch": 3.847199564980968,
      "grad_norm": 1.2734375,
      "learning_rate": 1.2190914318573895e-05,
      "loss": 0.0846,
      "step": 14150
    },
    {
      "epoch": 3.847199564980968,
      "eval_loss": 0.11810937523841858,
      "eval_runtime": 99.2203,
      "eval_samples_per_second": 5.039,
      "eval_steps_per_second": 0.161,
      "step": 14150
    },
    {
      "epoch": 3.8499184339314843,
      "grad_norm": 1.71875,
      "learning_rate": 1.2162162162162164e-05,
      "loss": 0.0964,
      "step": 14160
    },
    {
      "epoch": 3.852637302882001,
      "grad_norm": 0.51953125,
      "learning_rate": 1.2133410005750431e-05,
      "loss": 0.0866,
      "step": 14170
    },
    {
      "epoch": 3.8553561718325176,
      "grad_norm": 1.6328125,
      "learning_rate": 1.21046578493387e-05,
      "loss": 0.0975,
      "step": 14180
    },
    {
      "epoch": 3.858075040783034,
      "grad_norm": 1.8046875,
      "learning_rate": 1.207590569292697e-05,
      "loss": 0.1002,
      "step": 14190
    },
    {
      "epoch": 3.860793909733551,
      "grad_norm": 1.578125,
      "learning_rate": 1.204715353651524e-05,
      "loss": 0.0876,
      "step": 14200
    },
    {
      "epoch": 3.860793909733551,
      "eval_loss": 0.11759375035762787,
      "eval_runtime": 98.145,
      "eval_samples_per_second": 5.095,
      "eval_steps_per_second": 0.163,
      "step": 14200
    },
    {
      "epoch": 3.8635127786840675,
      "grad_norm": 3.53125,
      "learning_rate": 1.2018401380103509e-05,
      "loss": 0.1102,
      "step": 14210
    },
    {
      "epoch": 3.866231647634584,
      "grad_norm": 1.25,
      "learning_rate": 1.1989649223691778e-05,
      "loss": 0.0869,
      "step": 14220
    },
    {
      "epoch": 3.868950516585101,
      "grad_norm": 1.0234375,
      "learning_rate": 1.1960897067280047e-05,
      "loss": 0.0874,
      "step": 14230
    },
    {
      "epoch": 3.8716693855356175,
      "grad_norm": 1.9375,
      "learning_rate": 1.1932144910868316e-05,
      "loss": 0.0944,
      "step": 14240
    },
    {
      "epoch": 3.8743882544861337,
      "grad_norm": 1.015625,
      "learning_rate": 1.1903392754456585e-05,
      "loss": 0.0805,
      "step": 14250
    },
    {
      "epoch": 3.8743882544861337,
      "eval_loss": 0.11735156178474426,
      "eval_runtime": 97.0464,
      "eval_samples_per_second": 5.152,
      "eval_steps_per_second": 0.165,
      "step": 14250
    },
    {
      "epoch": 3.8771071234366503,
      "grad_norm": 0.796875,
      "learning_rate": 1.1874640598044854e-05,
      "loss": 0.0871,
      "step": 14260
    },
    {
      "epoch": 3.879825992387167,
      "grad_norm": 2.40625,
      "learning_rate": 1.1845888441633123e-05,
      "loss": 0.0841,
      "step": 14270
    },
    {
      "epoch": 3.8825448613376836,
      "grad_norm": 1.5234375,
      "learning_rate": 1.1817136285221392e-05,
      "loss": 0.0942,
      "step": 14280
    },
    {
      "epoch": 3.8852637302882003,
      "grad_norm": 0.8359375,
      "learning_rate": 1.178838412880966e-05,
      "loss": 0.0873,
      "step": 14290
    },
    {
      "epoch": 3.8879825992387165,
      "grad_norm": 0.921875,
      "learning_rate": 1.175963197239793e-05,
      "loss": 0.0742,
      "step": 14300
    },
    {
      "epoch": 3.8879825992387165,
      "eval_loss": 0.11278124898672104,
      "eval_runtime": 99.8446,
      "eval_samples_per_second": 5.008,
      "eval_steps_per_second": 0.16,
      "step": 14300
    },
    {
      "epoch": 3.890701468189233,
      "grad_norm": 1.3984375,
      "learning_rate": 1.17308798159862e-05,
      "loss": 0.0926,
      "step": 14310
    },
    {
      "epoch": 3.8934203371397498,
      "grad_norm": 1.7109375,
      "learning_rate": 1.170212765957447e-05,
      "loss": 0.0888,
      "step": 14320
    },
    {
      "epoch": 3.8961392060902664,
      "grad_norm": 1.6796875,
      "learning_rate": 1.1673375503162738e-05,
      "loss": 0.0919,
      "step": 14330
    },
    {
      "epoch": 3.898858075040783,
      "grad_norm": 1.734375,
      "learning_rate": 1.1644623346751007e-05,
      "loss": 0.0812,
      "step": 14340
    },
    {
      "epoch": 3.9015769439912997,
      "grad_norm": 0.89453125,
      "learning_rate": 1.1615871190339276e-05,
      "loss": 0.0879,
      "step": 14350
    },
    {
      "epoch": 3.9015769439912997,
      "eval_loss": 0.11561718583106995,
      "eval_runtime": 156.8985,
      "eval_samples_per_second": 3.187,
      "eval_steps_per_second": 0.102,
      "step": 14350
    },
    {
      "epoch": 3.9042958129418164,
      "grad_norm": 0.69921875,
      "learning_rate": 1.1587119033927544e-05,
      "loss": 0.105,
      "step": 14360
    },
    {
      "epoch": 3.907014681892333,
      "grad_norm": 2.390625,
      "learning_rate": 1.1558366877515814e-05,
      "loss": 0.085,
      "step": 14370
    },
    {
      "epoch": 3.9097335508428492,
      "grad_norm": 2.125,
      "learning_rate": 1.1529614721104083e-05,
      "loss": 0.0973,
      "step": 14380
    },
    {
      "epoch": 3.912452419793366,
      "grad_norm": 1.984375,
      "learning_rate": 1.1500862564692352e-05,
      "loss": 0.106,
      "step": 14390
    },
    {
      "epoch": 3.9151712887438825,
      "grad_norm": 1.2890625,
      "learning_rate": 1.1472110408280621e-05,
      "loss": 0.0969,
      "step": 14400
    },
    {
      "epoch": 3.9151712887438825,
      "eval_loss": 0.12182031571865082,
      "eval_runtime": 161.8821,
      "eval_samples_per_second": 3.089,
      "eval_steps_per_second": 0.099,
      "step": 14400
    },
    {
      "epoch": 3.917890157694399,
      "grad_norm": 1.1875,
      "learning_rate": 1.144335825186889e-05,
      "loss": 0.1138,
      "step": 14410
    },
    {
      "epoch": 3.920609026644916,
      "grad_norm": 3.25,
      "learning_rate": 1.141460609545716e-05,
      "loss": 0.0768,
      "step": 14420
    },
    {
      "epoch": 3.923327895595432,
      "grad_norm": 1.265625,
      "learning_rate": 1.138585393904543e-05,
      "loss": 0.0836,
      "step": 14430
    },
    {
      "epoch": 3.9260467645459487,
      "grad_norm": 1.484375,
      "learning_rate": 1.13571017826337e-05,
      "loss": 0.0828,
      "step": 14440
    },
    {
      "epoch": 3.9287656334964653,
      "grad_norm": 2.265625,
      "learning_rate": 1.1328349626221967e-05,
      "loss": 0.0943,
      "step": 14450
    },
    {
      "epoch": 3.9287656334964653,
      "eval_loss": 0.11989843845367432,
      "eval_runtime": 102.8984,
      "eval_samples_per_second": 4.859,
      "eval_steps_per_second": 0.155,
      "step": 14450
    },
    {
      "epoch": 3.931484502446982,
      "grad_norm": 1.671875,
      "learning_rate": 1.1299597469810236e-05,
      "loss": 0.0838,
      "step": 14460
    },
    {
      "epoch": 3.9342033713974986,
      "grad_norm": 1.984375,
      "learning_rate": 1.1270845313398505e-05,
      "loss": 0.0884,
      "step": 14470
    },
    {
      "epoch": 3.9369222403480153,
      "grad_norm": 0.94921875,
      "learning_rate": 1.1242093156986774e-05,
      "loss": 0.0915,
      "step": 14480
    },
    {
      "epoch": 3.939641109298532,
      "grad_norm": 1.9921875,
      "learning_rate": 1.1213341000575044e-05,
      "loss": 0.1032,
      "step": 14490
    },
    {
      "epoch": 3.9423599782490486,
      "grad_norm": 0.75390625,
      "learning_rate": 1.1184588844163313e-05,
      "loss": 0.0885,
      "step": 14500
    },
    {
      "epoch": 3.9423599782490486,
      "eval_loss": 0.1138906255364418,
      "eval_runtime": 97.8957,
      "eval_samples_per_second": 5.107,
      "eval_steps_per_second": 0.163,
      "step": 14500
    },
    {
      "epoch": 3.9450788471995653,
      "grad_norm": 1.78125,
      "learning_rate": 1.1155836687751582e-05,
      "loss": 0.0668,
      "step": 14510
    },
    {
      "epoch": 3.9477977161500815,
      "grad_norm": 1.390625,
      "learning_rate": 1.1127084531339851e-05,
      "loss": 0.093,
      "step": 14520
    },
    {
      "epoch": 3.950516585100598,
      "grad_norm": 1.1171875,
      "learning_rate": 1.109833237492812e-05,
      "loss": 0.0954,
      "step": 14530
    },
    {
      "epoch": 3.9532354540511148,
      "grad_norm": 1.5,
      "learning_rate": 1.106958021851639e-05,
      "loss": 0.0621,
      "step": 14540
    },
    {
      "epoch": 3.9559543230016314,
      "grad_norm": 0.98046875,
      "learning_rate": 1.1040828062104658e-05,
      "loss": 0.0843,
      "step": 14550
    },
    {
      "epoch": 3.9559543230016314,
      "eval_loss": 0.11607031524181366,
      "eval_runtime": 97.4011,
      "eval_samples_per_second": 5.133,
      "eval_steps_per_second": 0.164,
      "step": 14550
    },
    {
      "epoch": 3.958673191952148,
      "grad_norm": 1.6875,
      "learning_rate": 1.1012075905692927e-05,
      "loss": 0.0725,
      "step": 14560
    },
    {
      "epoch": 3.9613920609026643,
      "grad_norm": 1.8828125,
      "learning_rate": 1.0983323749281196e-05,
      "loss": 0.099,
      "step": 14570
    },
    {
      "epoch": 3.964110929853181,
      "grad_norm": 1.203125,
      "learning_rate": 1.0954571592869465e-05,
      "loss": 0.0833,
      "step": 14580
    },
    {
      "epoch": 3.9668297988036976,
      "grad_norm": 0.703125,
      "learning_rate": 1.0925819436457734e-05,
      "loss": 0.0866,
      "step": 14590
    },
    {
      "epoch": 3.969548667754214,
      "grad_norm": 1.1171875,
      "learning_rate": 1.0897067280046003e-05,
      "loss": 0.0762,
      "step": 14600
    },
    {
      "epoch": 3.969548667754214,
      "eval_loss": 0.12194531410932541,
      "eval_runtime": 96.7569,
      "eval_samples_per_second": 5.168,
      "eval_steps_per_second": 0.165,
      "step": 14600
    },
    {
      "epoch": 3.972267536704731,
      "grad_norm": 1.75,
      "learning_rate": 1.0868315123634274e-05,
      "loss": 0.0774,
      "step": 14610
    },
    {
      "epoch": 3.9749864056552475,
      "grad_norm": 1.0390625,
      "learning_rate": 1.0839562967222543e-05,
      "loss": 0.1187,
      "step": 14620
    },
    {
      "epoch": 3.977705274605764,
      "grad_norm": 1.5546875,
      "learning_rate": 1.0810810810810812e-05,
      "loss": 0.1233,
      "step": 14630
    },
    {
      "epoch": 3.980424143556281,
      "grad_norm": 1.6171875,
      "learning_rate": 1.078205865439908e-05,
      "loss": 0.0824,
      "step": 14640
    },
    {
      "epoch": 3.983143012506797,
      "grad_norm": 1.109375,
      "learning_rate": 1.0753306497987348e-05,
      "loss": 0.0963,
      "step": 14650
    },
    {
      "epoch": 3.983143012506797,
      "eval_loss": 0.11037500202655792,
      "eval_runtime": 97.9073,
      "eval_samples_per_second": 5.107,
      "eval_steps_per_second": 0.163,
      "step": 14650
    },
    {
      "epoch": 3.9858618814573137,
      "grad_norm": 2.125,
      "learning_rate": 1.0724554341575619e-05,
      "loss": 0.0871,
      "step": 14660
    },
    {
      "epoch": 3.9885807504078303,
      "grad_norm": 1.4609375,
      "learning_rate": 1.0695802185163888e-05,
      "loss": 0.0749,
      "step": 14670
    },
    {
      "epoch": 3.991299619358347,
      "grad_norm": 2.3125,
      "learning_rate": 1.0667050028752157e-05,
      "loss": 0.0879,
      "step": 14680
    },
    {
      "epoch": 3.9940184883088636,
      "grad_norm": 0.9296875,
      "learning_rate": 1.0638297872340426e-05,
      "loss": 0.0855,
      "step": 14690
    },
    {
      "epoch": 3.99673735725938,
      "grad_norm": 3.015625,
      "learning_rate": 1.0609545715928695e-05,
      "loss": 0.0958,
      "step": 14700
    },
    {
      "epoch": 3.99673735725938,
      "eval_loss": 0.11903125047683716,
      "eval_runtime": 96.4474,
      "eval_samples_per_second": 5.184,
      "eval_steps_per_second": 0.166,
      "step": 14700
    },
    {
      "epoch": 3.9994562262098965,
      "grad_norm": 0.76171875,
      "learning_rate": 1.0580793559516964e-05,
      "loss": 0.088,
      "step": 14710
    },
    {
      "epoch": 4.002175095160413,
      "grad_norm": 1.6171875,
      "learning_rate": 1.0552041403105233e-05,
      "loss": 0.1214,
      "step": 14720
    },
    {
      "epoch": 4.00489396411093,
      "grad_norm": 1.0234375,
      "learning_rate": 1.0523289246693504e-05,
      "loss": 0.0941,
      "step": 14730
    },
    {
      "epoch": 4.0076128330614464,
      "grad_norm": 0.82421875,
      "learning_rate": 1.0494537090281771e-05,
      "loss": 0.094,
      "step": 14740
    },
    {
      "epoch": 4.010331702011963,
      "grad_norm": 1.28125,
      "learning_rate": 1.046578493387004e-05,
      "loss": 0.0938,
      "step": 14750
    },
    {
      "epoch": 4.010331702011963,
      "eval_loss": 0.11144531518220901,
      "eval_runtime": 112.803,
      "eval_samples_per_second": 4.433,
      "eval_steps_per_second": 0.142,
      "step": 14750
    },
    {
      "epoch": 4.01305057096248,
      "grad_norm": 1.4765625,
      "learning_rate": 1.043703277745831e-05,
      "loss": 0.0898,
      "step": 14760
    },
    {
      "epoch": 4.015769439912996,
      "grad_norm": 1.21875,
      "learning_rate": 1.0408280621046578e-05,
      "loss": 0.0847,
      "step": 14770
    },
    {
      "epoch": 4.018488308863513,
      "grad_norm": 1.2734375,
      "learning_rate": 1.0379528464634849e-05,
      "loss": 0.0674,
      "step": 14780
    },
    {
      "epoch": 4.02120717781403,
      "grad_norm": 0.90234375,
      "learning_rate": 1.0350776308223118e-05,
      "loss": 0.0908,
      "step": 14790
    },
    {
      "epoch": 4.023926046764546,
      "grad_norm": 1.609375,
      "learning_rate": 1.0322024151811387e-05,
      "loss": 0.0965,
      "step": 14800
    },
    {
      "epoch": 4.023926046764546,
      "eval_loss": 0.11262500286102295,
      "eval_runtime": 97.2365,
      "eval_samples_per_second": 5.142,
      "eval_steps_per_second": 0.165,
      "step": 14800
    },
    {
      "epoch": 4.026644915715062,
      "grad_norm": 1.4921875,
      "learning_rate": 1.0293271995399656e-05,
      "loss": 0.0871,
      "step": 14810
    },
    {
      "epoch": 4.029363784665579,
      "grad_norm": 2.21875,
      "learning_rate": 1.0264519838987925e-05,
      "loss": 0.0708,
      "step": 14820
    },
    {
      "epoch": 4.032082653616095,
      "grad_norm": 2.125,
      "learning_rate": 1.0235767682576192e-05,
      "loss": 0.096,
      "step": 14830
    },
    {
      "epoch": 4.034801522566612,
      "grad_norm": 1.171875,
      "learning_rate": 1.0207015526164463e-05,
      "loss": 0.0828,
      "step": 14840
    },
    {
      "epoch": 4.037520391517129,
      "grad_norm": 2.75,
      "learning_rate": 1.0178263369752732e-05,
      "loss": 0.0895,
      "step": 14850
    },
    {
      "epoch": 4.037520391517129,
      "eval_loss": 0.11552343517541885,
      "eval_runtime": 117.2905,
      "eval_samples_per_second": 4.263,
      "eval_steps_per_second": 0.136,
      "step": 14850
    },
    {
      "epoch": 4.040239260467645,
      "grad_norm": 2.953125,
      "learning_rate": 1.0149511213341001e-05,
      "loss": 0.0833,
      "step": 14860
    },
    {
      "epoch": 4.042958129418162,
      "grad_norm": 3.09375,
      "learning_rate": 1.012075905692927e-05,
      "loss": 0.0849,
      "step": 14870
    },
    {
      "epoch": 4.045676998368679,
      "grad_norm": 1.1328125,
      "learning_rate": 1.0092006900517539e-05,
      "loss": 0.0839,
      "step": 14880
    },
    {
      "epoch": 4.048395867319195,
      "grad_norm": 1.3359375,
      "learning_rate": 1.0063254744105808e-05,
      "loss": 0.0825,
      "step": 14890
    },
    {
      "epoch": 4.051114736269712,
      "grad_norm": 1.4765625,
      "learning_rate": 1.0034502587694079e-05,
      "loss": 0.1007,
      "step": 14900
    },
    {
      "epoch": 4.051114736269712,
      "eval_loss": 0.11796093732118607,
      "eval_runtime": 102.6959,
      "eval_samples_per_second": 4.869,
      "eval_steps_per_second": 0.156,
      "step": 14900
    },
    {
      "epoch": 4.053833605220229,
      "grad_norm": 1.0234375,
      "learning_rate": 1.0005750431282348e-05,
      "loss": 0.076,
      "step": 14910
    },
    {
      "epoch": 4.056552474170745,
      "grad_norm": 2.15625,
      "learning_rate": 9.976998274870617e-06,
      "loss": 0.0933,
      "step": 14920
    },
    {
      "epoch": 4.059271343121262,
      "grad_norm": 1.53125,
      "learning_rate": 9.948246118458884e-06,
      "loss": 0.1021,
      "step": 14930
    },
    {
      "epoch": 4.061990212071779,
      "grad_norm": 3.453125,
      "learning_rate": 9.919493962047153e-06,
      "loss": 0.0932,
      "step": 14940
    },
    {
      "epoch": 4.064709081022294,
      "grad_norm": 1.90625,
      "learning_rate": 9.890741805635422e-06,
      "loss": 0.0939,
      "step": 14950
    },
    {
      "epoch": 4.064709081022294,
      "eval_loss": 0.1103515625,
      "eval_runtime": 100.5532,
      "eval_samples_per_second": 4.972,
      "eval_steps_per_second": 0.159,
      "step": 14950
    },
    {
      "epoch": 4.067427949972811,
      "grad_norm": 1.40625,
      "learning_rate": 9.861989649223693e-06,
      "loss": 0.0738,
      "step": 14960
    },
    {
      "epoch": 4.070146818923328,
      "grad_norm": 1.3828125,
      "learning_rate": 9.833237492811962e-06,
      "loss": 0.099,
      "step": 14970
    },
    {
      "epoch": 4.072865687873844,
      "grad_norm": 1.4375,
      "learning_rate": 9.80448533640023e-06,
      "loss": 0.0681,
      "step": 14980
    },
    {
      "epoch": 4.075584556824361,
      "grad_norm": 1.5390625,
      "learning_rate": 9.7757331799885e-06,
      "loss": 0.0707,
      "step": 14990
    },
    {
      "epoch": 4.078303425774878,
      "grad_norm": 1.2109375,
      "learning_rate": 9.746981023576769e-06,
      "loss": 0.1142,
      "step": 15000
    },
    {
      "epoch": 4.078303425774878,
      "eval_loss": 0.10621874779462814,
      "eval_runtime": 99.8531,
      "eval_samples_per_second": 5.007,
      "eval_steps_per_second": 0.16,
      "step": 15000
    },
    {
      "epoch": 4.081022294725394,
      "grad_norm": 0.84765625,
      "learning_rate": 9.718228867165038e-06,
      "loss": 0.067,
      "step": 15010
    },
    {
      "epoch": 4.083741163675911,
      "grad_norm": 1.6015625,
      "learning_rate": 9.689476710753307e-06,
      "loss": 0.087,
      "step": 15020
    },
    {
      "epoch": 4.0864600326264275,
      "grad_norm": 1.3984375,
      "learning_rate": 9.660724554341576e-06,
      "loss": 0.0925,
      "step": 15030
    },
    {
      "epoch": 4.089178901576944,
      "grad_norm": 2.046875,
      "learning_rate": 9.631972397929845e-06,
      "loss": 0.0879,
      "step": 15040
    },
    {
      "epoch": 4.091897770527461,
      "grad_norm": 2.796875,
      "learning_rate": 9.603220241518114e-06,
      "loss": 0.0871,
      "step": 15050
    },
    {
      "epoch": 4.091897770527461,
      "eval_loss": 0.11582031100988388,
      "eval_runtime": 98.5669,
      "eval_samples_per_second": 5.073,
      "eval_steps_per_second": 0.162,
      "step": 15050
    },
    {
      "epoch": 4.0946166394779775,
      "grad_norm": 1.375,
      "learning_rate": 9.574468085106383e-06,
      "loss": 0.0894,
      "step": 15060
    },
    {
      "epoch": 4.097335508428494,
      "grad_norm": 1.515625,
      "learning_rate": 9.545715928694652e-06,
      "loss": 0.0604,
      "step": 15070
    },
    {
      "epoch": 4.10005437737901,
      "grad_norm": 1.4609375,
      "learning_rate": 9.516963772282923e-06,
      "loss": 0.0644,
      "step": 15080
    },
    {
      "epoch": 4.102773246329527,
      "grad_norm": 1.3125,
      "learning_rate": 9.488211615871192e-06,
      "loss": 0.0761,
      "step": 15090
    },
    {
      "epoch": 4.105492115280043,
      "grad_norm": 1.3515625,
      "learning_rate": 9.45945945945946e-06,
      "loss": 0.0787,
      "step": 15100
    },
    {
      "epoch": 4.105492115280043,
      "eval_loss": 0.11356250196695328,
      "eval_runtime": 99.2038,
      "eval_samples_per_second": 5.04,
      "eval_steps_per_second": 0.161,
      "step": 15100
    },
    {
      "epoch": 4.10821098423056,
      "grad_norm": 1.7578125,
      "learning_rate": 9.43070730304773e-06,
      "loss": 0.1094,
      "step": 15110
    },
    {
      "epoch": 4.1109298531810765,
      "grad_norm": 1.7734375,
      "learning_rate": 9.401955146635997e-06,
      "loss": 0.0839,
      "step": 15120
    },
    {
      "epoch": 4.113648722131593,
      "grad_norm": 0.95703125,
      "learning_rate": 9.373202990224268e-06,
      "loss": 0.0855,
      "step": 15130
    },
    {
      "epoch": 4.11636759108211,
      "grad_norm": 2.21875,
      "learning_rate": 9.344450833812537e-06,
      "loss": 0.0949,
      "step": 15140
    },
    {
      "epoch": 4.1190864600326265,
      "grad_norm": 1.2421875,
      "learning_rate": 9.315698677400806e-06,
      "loss": 0.0876,
      "step": 15150
    },
    {
      "epoch": 4.1190864600326265,
      "eval_loss": 0.1113593727350235,
      "eval_runtime": 97.9755,
      "eval_samples_per_second": 5.103,
      "eval_steps_per_second": 0.163,
      "step": 15150
    },
    {
      "epoch": 4.121805328983143,
      "grad_norm": 1.03125,
      "learning_rate": 9.286946520989075e-06,
      "loss": 0.1136,
      "step": 15160
    },
    {
      "epoch": 4.12452419793366,
      "grad_norm": 1.734375,
      "learning_rate": 9.258194364577344e-06,
      "loss": 0.0849,
      "step": 15170
    },
    {
      "epoch": 4.127243066884176,
      "grad_norm": 0.73046875,
      "learning_rate": 9.229442208165613e-06,
      "loss": 0.0929,
      "step": 15180
    },
    {
      "epoch": 4.129961935834693,
      "grad_norm": 0.9140625,
      "learning_rate": 9.200690051753882e-06,
      "loss": 0.0957,
      "step": 15190
    },
    {
      "epoch": 4.13268080478521,
      "grad_norm": 1.2890625,
      "learning_rate": 9.171937895342152e-06,
      "loss": 0.0863,
      "step": 15200
    },
    {
      "epoch": 4.13268080478521,
      "eval_loss": 0.12992969155311584,
      "eval_runtime": 100.5068,
      "eval_samples_per_second": 4.975,
      "eval_steps_per_second": 0.159,
      "step": 15200
    },
    {
      "epoch": 4.135399673735726,
      "grad_norm": 2.96875,
      "learning_rate": 9.14318573893042e-06,
      "loss": 0.0883,
      "step": 15210
    },
    {
      "epoch": 4.138118542686242,
      "grad_norm": 2.15625,
      "learning_rate": 9.114433582518689e-06,
      "loss": 0.0763,
      "step": 15220
    },
    {
      "epoch": 4.140837411636759,
      "grad_norm": 0.95703125,
      "learning_rate": 9.085681426106958e-06,
      "loss": 0.0666,
      "step": 15230
    },
    {
      "epoch": 4.143556280587275,
      "grad_norm": 0.90625,
      "learning_rate": 9.056929269695227e-06,
      "loss": 0.0824,
      "step": 15240
    },
    {
      "epoch": 4.146275149537792,
      "grad_norm": 1.6953125,
      "learning_rate": 9.028177113283497e-06,
      "loss": 0.0848,
      "step": 15250
    },
    {
      "epoch": 4.146275149537792,
      "eval_loss": 0.10918749868869781,
      "eval_runtime": 97.7341,
      "eval_samples_per_second": 5.116,
      "eval_steps_per_second": 0.164,
      "step": 15250
    },
    {
      "epoch": 4.148994018488309,
      "grad_norm": 2.359375,
      "learning_rate": 8.999424956871766e-06,
      "loss": 0.0657,
      "step": 15260
    },
    {
      "epoch": 4.151712887438825,
      "grad_norm": 1.4921875,
      "learning_rate": 8.970672800460035e-06,
      "loss": 0.0674,
      "step": 15270
    },
    {
      "epoch": 4.154431756389342,
      "grad_norm": 1.3828125,
      "learning_rate": 8.941920644048304e-06,
      "loss": 0.0815,
      "step": 15280
    },
    {
      "epoch": 4.157150625339859,
      "grad_norm": 1.9375,
      "learning_rate": 8.913168487636573e-06,
      "loss": 0.1098,
      "step": 15290
    },
    {
      "epoch": 4.159869494290375,
      "grad_norm": 2.046875,
      "learning_rate": 8.884416331224842e-06,
      "loss": 0.0738,
      "step": 15300
    },
    {
      "epoch": 4.159869494290375,
      "eval_loss": 0.11103124916553497,
      "eval_runtime": 97.7309,
      "eval_samples_per_second": 5.116,
      "eval_steps_per_second": 0.164,
      "step": 15300
    },
    {
      "epoch": 4.162588363240892,
      "grad_norm": 2.65625,
      "learning_rate": 8.855664174813111e-06,
      "loss": 0.0989,
      "step": 15310
    },
    {
      "epoch": 4.165307232191409,
      "grad_norm": 1.5546875,
      "learning_rate": 8.82691201840138e-06,
      "loss": 0.0827,
      "step": 15320
    },
    {
      "epoch": 4.168026101141925,
      "grad_norm": 2.28125,
      "learning_rate": 8.79815986198965e-06,
      "loss": 0.0858,
      "step": 15330
    },
    {
      "epoch": 4.170744970092442,
      "grad_norm": 1.59375,
      "learning_rate": 8.769407705577918e-06,
      "loss": 0.0759,
      "step": 15340
    },
    {
      "epoch": 4.173463839042958,
      "grad_norm": 0.94140625,
      "learning_rate": 8.740655549166187e-06,
      "loss": 0.0711,
      "step": 15350
    },
    {
      "epoch": 4.173463839042958,
      "eval_loss": 0.1139531284570694,
      "eval_runtime": 125.9068,
      "eval_samples_per_second": 3.971,
      "eval_steps_per_second": 0.127,
      "step": 15350
    },
    {
      "epoch": 4.176182707993474,
      "grad_norm": 1.578125,
      "learning_rate": 8.711903392754456e-06,
      "loss": 0.0861,
      "step": 15360
    },
    {
      "epoch": 4.178901576943991,
      "grad_norm": 1.71875,
      "learning_rate": 8.683151236342727e-06,
      "loss": 0.0933,
      "step": 15370
    },
    {
      "epoch": 4.181620445894508,
      "grad_norm": 0.91796875,
      "learning_rate": 8.654399079930996e-06,
      "loss": 0.0805,
      "step": 15380
    },
    {
      "epoch": 4.184339314845024,
      "grad_norm": 1.15625,
      "learning_rate": 8.625646923519265e-06,
      "loss": 0.0892,
      "step": 15390
    },
    {
      "epoch": 4.187058183795541,
      "grad_norm": 0.77734375,
      "learning_rate": 8.596894767107533e-06,
      "loss": 0.0936,
      "step": 15400
    },
    {
      "epoch": 4.187058183795541,
      "eval_loss": 0.10895312577486038,
      "eval_runtime": 104.4641,
      "eval_samples_per_second": 4.786,
      "eval_steps_per_second": 0.153,
      "step": 15400
    },
    {
      "epoch": 4.189777052746058,
      "grad_norm": 1.7421875,
      "learning_rate": 8.568142610695802e-06,
      "loss": 0.0864,
      "step": 15410
    },
    {
      "epoch": 4.192495921696574,
      "grad_norm": 1.4296875,
      "learning_rate": 8.53939045428407e-06,
      "loss": 0.0794,
      "step": 15420
    },
    {
      "epoch": 4.195214790647091,
      "grad_norm": 0.640625,
      "learning_rate": 8.510638297872341e-06,
      "loss": 0.0725,
      "step": 15430
    },
    {
      "epoch": 4.197933659597608,
      "grad_norm": 1.5234375,
      "learning_rate": 8.48188614146061e-06,
      "loss": 0.0812,
      "step": 15440
    },
    {
      "epoch": 4.200652528548124,
      "grad_norm": 2.34375,
      "learning_rate": 8.45313398504888e-06,
      "loss": 0.0779,
      "step": 15450
    },
    {
      "epoch": 4.200652528548124,
      "eval_loss": 0.10947655886411667,
      "eval_runtime": 99.793,
      "eval_samples_per_second": 5.01,
      "eval_steps_per_second": 0.16,
      "step": 15450
    },
    {
      "epoch": 4.203371397498641,
      "grad_norm": 1.0625,
      "learning_rate": 8.424381828637148e-06,
      "loss": 0.0593,
      "step": 15460
    },
    {
      "epoch": 4.2060902664491575,
      "grad_norm": 1.0390625,
      "learning_rate": 8.395629672225417e-06,
      "loss": 0.0953,
      "step": 15470
    },
    {
      "epoch": 4.208809135399674,
      "grad_norm": 1.0625,
      "learning_rate": 8.366877515813686e-06,
      "loss": 0.0959,
      "step": 15480
    },
    {
      "epoch": 4.21152800435019,
      "grad_norm": 1.4375,
      "learning_rate": 8.338125359401957e-06,
      "loss": 0.072,
      "step": 15490
    },
    {
      "epoch": 4.214246873300707,
      "grad_norm": 1.421875,
      "learning_rate": 8.309373202990224e-06,
      "loss": 0.0778,
      "step": 15500
    },
    {
      "epoch": 4.214246873300707,
      "eval_loss": 0.11453906446695328,
      "eval_runtime": 117.1704,
      "eval_samples_per_second": 4.267,
      "eval_steps_per_second": 0.137,
      "step": 15500
    },
    {
      "epoch": 4.216965742251223,
      "grad_norm": 1.4296875,
      "learning_rate": 8.280621046578493e-06,
      "loss": 0.0883,
      "step": 15510
    },
    {
      "epoch": 4.21968461120174,
      "grad_norm": 1.2421875,
      "learning_rate": 8.251868890166762e-06,
      "loss": 0.0765,
      "step": 15520
    },
    {
      "epoch": 4.2224034801522565,
      "grad_norm": 2.5,
      "learning_rate": 8.223116733755031e-06,
      "loss": 0.1029,
      "step": 15530
    },
    {
      "epoch": 4.225122349102773,
      "grad_norm": 1.171875,
      "learning_rate": 8.1943645773433e-06,
      "loss": 0.0604,
      "step": 15540
    },
    {
      "epoch": 4.22784121805329,
      "grad_norm": 1.859375,
      "learning_rate": 8.165612420931571e-06,
      "loss": 0.0829,
      "step": 15550
    },
    {
      "epoch": 4.22784121805329,
      "eval_loss": 0.11235155910253525,
      "eval_runtime": 114.5546,
      "eval_samples_per_second": 4.365,
      "eval_steps_per_second": 0.14,
      "step": 15550
    },
    {
      "epoch": 4.2305600870038065,
      "grad_norm": 0.80859375,
      "learning_rate": 8.13686026451984e-06,
      "loss": 0.0765,
      "step": 15560
    },
    {
      "epoch": 4.233278955954323,
      "grad_norm": 2.15625,
      "learning_rate": 8.108108108108109e-06,
      "loss": 0.0954,
      "step": 15570
    },
    {
      "epoch": 4.23599782490484,
      "grad_norm": 1.03125,
      "learning_rate": 8.079355951696378e-06,
      "loss": 0.1007,
      "step": 15580
    },
    {
      "epoch": 4.238716693855356,
      "grad_norm": 1.40625,
      "learning_rate": 8.050603795284645e-06,
      "loss": 0.0851,
      "step": 15590
    },
    {
      "epoch": 4.241435562805873,
      "grad_norm": 0.640625,
      "learning_rate": 8.021851638872916e-06,
      "loss": 0.1039,
      "step": 15600
    },
    {
      "epoch": 4.241435562805873,
      "eval_loss": 0.11228124797344208,
      "eval_runtime": 99.7342,
      "eval_samples_per_second": 5.013,
      "eval_steps_per_second": 0.16,
      "step": 15600
    },
    {
      "epoch": 4.24415443175639,
      "grad_norm": 1.6328125,
      "learning_rate": 7.993099482461185e-06,
      "loss": 0.0843,
      "step": 15610
    },
    {
      "epoch": 4.2468733007069055,
      "grad_norm": 0.59375,
      "learning_rate": 7.964347326049454e-06,
      "loss": 0.0805,
      "step": 15620
    },
    {
      "epoch": 4.249592169657422,
      "grad_norm": 1.125,
      "learning_rate": 7.935595169637723e-06,
      "loss": 0.069,
      "step": 15630
    },
    {
      "epoch": 4.252311038607939,
      "grad_norm": 1.46875,
      "learning_rate": 7.906843013225992e-06,
      "loss": 0.0717,
      "step": 15640
    },
    {
      "epoch": 4.2550299075584554,
      "grad_norm": 2.046875,
      "learning_rate": 7.878090856814261e-06,
      "loss": 0.0874,
      "step": 15650
    },
    {
      "epoch": 4.2550299075584554,
      "eval_loss": 0.11900781095027924,
      "eval_runtime": 114.6497,
      "eval_samples_per_second": 4.361,
      "eval_steps_per_second": 0.14,
      "step": 15650
    },
    {
      "epoch": 4.257748776508972,
      "grad_norm": 1.875,
      "learning_rate": 7.84933870040253e-06,
      "loss": 0.0787,
      "step": 15660
    },
    {
      "epoch": 4.260467645459489,
      "grad_norm": 1.359375,
      "learning_rate": 7.8205865439908e-06,
      "loss": 0.0802,
      "step": 15670
    },
    {
      "epoch": 4.263186514410005,
      "grad_norm": 1.640625,
      "learning_rate": 7.79183438757907e-06,
      "loss": 0.0844,
      "step": 15680
    },
    {
      "epoch": 4.265905383360522,
      "grad_norm": 1.71875,
      "learning_rate": 7.763082231167337e-06,
      "loss": 0.1043,
      "step": 15690
    },
    {
      "epoch": 4.268624252311039,
      "grad_norm": 2.28125,
      "learning_rate": 7.734330074755606e-06,
      "loss": 0.0888,
      "step": 15700
    },
    {
      "epoch": 4.268624252311039,
      "eval_loss": 0.12896093726158142,
      "eval_runtime": 100.8507,
      "eval_samples_per_second": 4.958,
      "eval_steps_per_second": 0.159,
      "step": 15700
    },
    {
      "epoch": 4.271343121261555,
      "grad_norm": 1.5234375,
      "learning_rate": 7.705577918343875e-06,
      "loss": 0.0873,
      "step": 15710
    },
    {
      "epoch": 4.274061990212072,
      "grad_norm": 0.7109375,
      "learning_rate": 7.676825761932146e-06,
      "loss": 0.0831,
      "step": 15720
    },
    {
      "epoch": 4.276780859162589,
      "grad_norm": 1.1015625,
      "learning_rate": 7.648073605520415e-06,
      "loss": 0.0816,
      "step": 15730
    },
    {
      "epoch": 4.279499728113105,
      "grad_norm": 0.921875,
      "learning_rate": 7.619321449108684e-06,
      "loss": 0.0771,
      "step": 15740
    },
    {
      "epoch": 4.282218597063622,
      "grad_norm": 1.8203125,
      "learning_rate": 7.590569292696953e-06,
      "loss": 0.1031,
      "step": 15750
    },
    {
      "epoch": 4.282218597063622,
      "eval_loss": 0.11036718636751175,
      "eval_runtime": 100.4928,
      "eval_samples_per_second": 4.975,
      "eval_steps_per_second": 0.159,
      "step": 15750
    },
    {
      "epoch": 4.284937466014138,
      "grad_norm": 1.8359375,
      "learning_rate": 7.561817136285222e-06,
      "loss": 0.0945,
      "step": 15760
    },
    {
      "epoch": 4.287656334964654,
      "grad_norm": 1.34375,
      "learning_rate": 7.533064979873492e-06,
      "loss": 0.0752,
      "step": 15770
    },
    {
      "epoch": 4.290375203915171,
      "grad_norm": 1.8046875,
      "learning_rate": 7.504312823461759e-06,
      "loss": 0.0842,
      "step": 15780
    },
    {
      "epoch": 4.293094072865688,
      "grad_norm": 2.6875,
      "learning_rate": 7.475560667050029e-06,
      "loss": 0.0741,
      "step": 15790
    },
    {
      "epoch": 4.295812941816204,
      "grad_norm": 1.6640625,
      "learning_rate": 7.446808510638298e-06,
      "loss": 0.0715,
      "step": 15800
    },
    {
      "epoch": 4.295812941816204,
      "eval_loss": 0.11823437362909317,
      "eval_runtime": 98.7714,
      "eval_samples_per_second": 5.062,
      "eval_steps_per_second": 0.162,
      "step": 15800
    },
    {
      "epoch": 4.298531810766721,
      "grad_norm": 1.140625,
      "learning_rate": 7.418056354226567e-06,
      "loss": 0.0896,
      "step": 15810
    },
    {
      "epoch": 4.301250679717238,
      "grad_norm": 1.96875,
      "learning_rate": 7.389304197814837e-06,
      "loss": 0.0802,
      "step": 15820
    },
    {
      "epoch": 4.303969548667754,
      "grad_norm": 1.0078125,
      "learning_rate": 7.360552041403106e-06,
      "loss": 0.0852,
      "step": 15830
    },
    {
      "epoch": 4.306688417618271,
      "grad_norm": 1.65625,
      "learning_rate": 7.331799884991375e-06,
      "loss": 0.1011,
      "step": 15840
    },
    {
      "epoch": 4.309407286568788,
      "grad_norm": 1.546875,
      "learning_rate": 7.303047728579644e-06,
      "loss": 0.0953,
      "step": 15850
    },
    {
      "epoch": 4.309407286568788,
      "eval_loss": 0.12019531428813934,
      "eval_runtime": 111.4698,
      "eval_samples_per_second": 4.486,
      "eval_steps_per_second": 0.144,
      "step": 15850
    },
    {
      "epoch": 4.312126155519304,
      "grad_norm": 1.671875,
      "learning_rate": 7.274295572167914e-06,
      "loss": 0.0823,
      "step": 15860
    },
    {
      "epoch": 4.314845024469821,
      "grad_norm": 1.6015625,
      "learning_rate": 7.245543415756183e-06,
      "loss": 0.0934,
      "step": 15870
    },
    {
      "epoch": 4.3175638934203375,
      "grad_norm": 1.4765625,
      "learning_rate": 7.216791259344451e-06,
      "loss": 0.0843,
      "step": 15880
    },
    {
      "epoch": 4.320282762370853,
      "grad_norm": 1.4453125,
      "learning_rate": 7.18803910293272e-06,
      "loss": 0.0722,
      "step": 15890
    },
    {
      "epoch": 4.32300163132137,
      "grad_norm": 2.015625,
      "learning_rate": 7.159286946520989e-06,
      "loss": 0.0829,
      "step": 15900
    },
    {
      "epoch": 4.32300163132137,
      "eval_loss": 0.11312499642372131,
      "eval_runtime": 102.6748,
      "eval_samples_per_second": 4.87,
      "eval_steps_per_second": 0.156,
      "step": 15900
    },
    {
      "epoch": 4.325720500271887,
      "grad_norm": 0.71484375,
      "learning_rate": 7.130534790109259e-06,
      "loss": 0.0912,
      "step": 15910
    },
    {
      "epoch": 4.328439369222403,
      "grad_norm": 1.7734375,
      "learning_rate": 7.101782633697528e-06,
      "loss": 0.0853,
      "step": 15920
    },
    {
      "epoch": 4.33115823817292,
      "grad_norm": 1.1640625,
      "learning_rate": 7.073030477285797e-06,
      "loss": 0.0877,
      "step": 15930
    },
    {
      "epoch": 4.3338771071234365,
      "grad_norm": 1.8359375,
      "learning_rate": 7.044278320874067e-06,
      "loss": 0.0901,
      "step": 15940
    },
    {
      "epoch": 4.336595976073953,
      "grad_norm": 1.6796875,
      "learning_rate": 7.015526164462336e-06,
      "loss": 0.0807,
      "step": 15950
    },
    {
      "epoch": 4.336595976073953,
      "eval_loss": 0.11236327886581421,
      "eval_runtime": 99.1879,
      "eval_samples_per_second": 5.041,
      "eval_steps_per_second": 0.161,
      "step": 15950
    },
    {
      "epoch": 4.33931484502447,
      "grad_norm": 2.53125,
      "learning_rate": 6.986774008050605e-06,
      "loss": 0.1066,
      "step": 15960
    },
    {
      "epoch": 4.3420337139749865,
      "grad_norm": 2.03125,
      "learning_rate": 6.958021851638873e-06,
      "loss": 0.0752,
      "step": 15970
    },
    {
      "epoch": 4.344752582925503,
      "grad_norm": 1.5234375,
      "learning_rate": 6.929269695227142e-06,
      "loss": 0.0687,
      "step": 15980
    },
    {
      "epoch": 4.34747145187602,
      "grad_norm": 1.1484375,
      "learning_rate": 6.900517538815411e-06,
      "loss": 0.0834,
      "step": 15990
    },
    {
      "epoch": 4.3501903208265364,
      "grad_norm": 1.9453125,
      "learning_rate": 6.871765382403681e-06,
      "loss": 0.0845,
      "step": 16000
    },
    {
      "epoch": 4.3501903208265364,
      "eval_loss": 0.11637499928474426,
      "eval_runtime": 99.4434,
      "eval_samples_per_second": 5.028,
      "eval_steps_per_second": 0.161,
      "step": 16000
    },
    {
      "epoch": 4.352909189777053,
      "grad_norm": 1.2265625,
      "learning_rate": 6.84301322599195e-06,
      "loss": 0.0752,
      "step": 16010
    },
    {
      "epoch": 4.35562805872757,
      "grad_norm": 0.98828125,
      "learning_rate": 6.814261069580219e-06,
      "loss": 0.0831,
      "step": 16020
    },
    {
      "epoch": 4.3583469276780855,
      "grad_norm": 2.078125,
      "learning_rate": 6.7855089131684885e-06,
      "loss": 0.0741,
      "step": 16030
    },
    {
      "epoch": 4.361065796628602,
      "grad_norm": 1.6953125,
      "learning_rate": 6.7567567567567575e-06,
      "loss": 0.0883,
      "step": 16040
    },
    {
      "epoch": 4.363784665579119,
      "grad_norm": 1.0859375,
      "learning_rate": 6.7280046003450265e-06,
      "loss": 0.0696,
      "step": 16050
    },
    {
      "epoch": 4.363784665579119,
      "eval_loss": 0.11478125303983688,
      "eval_runtime": 98.9199,
      "eval_samples_per_second": 5.055,
      "eval_steps_per_second": 0.162,
      "step": 16050
    },
    {
      "epoch": 4.3665035345296355,
      "grad_norm": 2.046875,
      "learning_rate": 6.699252443933296e-06,
      "loss": 0.0958,
      "step": 16060
    },
    {
      "epoch": 4.369222403480152,
      "grad_norm": 1.1953125,
      "learning_rate": 6.670500287521564e-06,
      "loss": 0.0777,
      "step": 16070
    },
    {
      "epoch": 4.371941272430669,
      "grad_norm": 2.1875,
      "learning_rate": 6.641748131109833e-06,
      "loss": 0.0836,
      "step": 16080
    },
    {
      "epoch": 4.374660141381185,
      "grad_norm": 1.1953125,
      "learning_rate": 6.612995974698103e-06,
      "loss": 0.0833,
      "step": 16090
    },
    {
      "epoch": 4.377379010331702,
      "grad_norm": 1.0546875,
      "learning_rate": 6.584243818286372e-06,
      "loss": 0.0808,
      "step": 16100
    },
    {
      "epoch": 4.377379010331702,
      "eval_loss": 0.1145937517285347,
      "eval_runtime": 100.0028,
      "eval_samples_per_second": 5.0,
      "eval_steps_per_second": 0.16,
      "step": 16100
    },
    {
      "epoch": 4.380097879282219,
      "grad_norm": 2.1875,
      "learning_rate": 6.555491661874641e-06,
      "loss": 0.0871,
      "step": 16110
    },
    {
      "epoch": 4.382816748232735,
      "grad_norm": 2.640625,
      "learning_rate": 6.5267395054629105e-06,
      "loss": 0.0783,
      "step": 16120
    },
    {
      "epoch": 4.385535617183252,
      "grad_norm": 1.1171875,
      "learning_rate": 6.4979873490511795e-06,
      "loss": 0.0883,
      "step": 16130
    },
    {
      "epoch": 4.388254486133769,
      "grad_norm": 1.2890625,
      "learning_rate": 6.4692351926394485e-06,
      "loss": 0.078,
      "step": 16140
    },
    {
      "epoch": 4.390973355084285,
      "grad_norm": 1.515625,
      "learning_rate": 6.440483036227718e-06,
      "loss": 0.0713,
      "step": 16150
    },
    {
      "epoch": 4.390973355084285,
      "eval_loss": 0.11408593505620956,
      "eval_runtime": 98.8091,
      "eval_samples_per_second": 5.06,
      "eval_steps_per_second": 0.162,
      "step": 16150
    },
    {
      "epoch": 4.393692224034801,
      "grad_norm": 1.390625,
      "learning_rate": 6.411730879815986e-06,
      "loss": 0.1074,
      "step": 16160
    },
    {
      "epoch": 4.396411092985318,
      "grad_norm": 1.4921875,
      "learning_rate": 6.3829787234042555e-06,
      "loss": 0.1002,
      "step": 16170
    },
    {
      "epoch": 4.399129961935834,
      "grad_norm": 1.625,
      "learning_rate": 6.3542265669925245e-06,
      "loss": 0.0718,
      "step": 16180
    },
    {
      "epoch": 4.401848830886351,
      "grad_norm": 1.828125,
      "learning_rate": 6.3254744105807935e-06,
      "loss": 0.0855,
      "step": 16190
    },
    {
      "epoch": 4.404567699836868,
      "grad_norm": 2.484375,
      "learning_rate": 6.2967222541690625e-06,
      "loss": 0.1083,
      "step": 16200
    },
    {
      "epoch": 4.404567699836868,
      "eval_loss": 0.1187734380364418,
      "eval_runtime": 99.6826,
      "eval_samples_per_second": 5.016,
      "eval_steps_per_second": 0.161,
      "step": 16200
    },
    {
      "epoch": 4.407286568787384,
      "grad_norm": 2.625,
      "learning_rate": 6.267970097757332e-06,
      "loss": 0.0911,
      "step": 16210
    },
    {
      "epoch": 4.410005437737901,
      "grad_norm": 1.4140625,
      "learning_rate": 6.239217941345601e-06,
      "loss": 0.0597,
      "step": 16220
    },
    {
      "epoch": 4.412724306688418,
      "grad_norm": 1.3046875,
      "learning_rate": 6.21046578493387e-06,
      "loss": 0.0828,
      "step": 16230
    },
    {
      "epoch": 4.415443175638934,
      "grad_norm": 1.5390625,
      "learning_rate": 6.181713628522139e-06,
      "loss": 0.09,
      "step": 16240
    },
    {
      "epoch": 4.418162044589451,
      "grad_norm": 1.8984375,
      "learning_rate": 6.152961472110408e-06,
      "loss": 0.1039,
      "step": 16250
    },
    {
      "epoch": 4.418162044589451,
      "eval_loss": 0.11271875351667404,
      "eval_runtime": 108.3281,
      "eval_samples_per_second": 4.616,
      "eval_steps_per_second": 0.148,
      "step": 16250
    },
    {
      "epoch": 4.420880913539968,
      "grad_norm": 1.171875,
      "learning_rate": 6.1242093156986774e-06,
      "loss": 0.066,
      "step": 16260
    },
    {
      "epoch": 4.423599782490484,
      "grad_norm": 2.578125,
      "learning_rate": 6.095457159286947e-06,
      "loss": 0.0834,
      "step": 16270
    },
    {
      "epoch": 4.426318651441001,
      "grad_norm": 0.74609375,
      "learning_rate": 6.0667050028752154e-06,
      "loss": 0.0805,
      "step": 16280
    },
    {
      "epoch": 4.4290375203915175,
      "grad_norm": 1.140625,
      "learning_rate": 6.037952846463485e-06,
      "loss": 0.0854,
      "step": 16290
    },
    {
      "epoch": 4.431756389342033,
      "grad_norm": 1.765625,
      "learning_rate": 6.009200690051754e-06,
      "loss": 0.0708,
      "step": 16300
    },
    {
      "epoch": 4.431756389342033,
      "eval_loss": 0.1138906255364418,
      "eval_runtime": 97.6798,
      "eval_samples_per_second": 5.119,
      "eval_steps_per_second": 0.164,
      "step": 16300
    },
    {
      "epoch": 4.43447525829255,
      "grad_norm": 1.2734375,
      "learning_rate": 5.980448533640023e-06,
      "loss": 0.0841,
      "step": 16310
    },
    {
      "epoch": 4.437194127243067,
      "grad_norm": 2.015625,
      "learning_rate": 5.951696377228292e-06,
      "loss": 0.0759,
      "step": 16320
    },
    {
      "epoch": 4.439912996193583,
      "grad_norm": 1.546875,
      "learning_rate": 5.922944220816561e-06,
      "loss": 0.0637,
      "step": 16330
    },
    {
      "epoch": 4.4426318651441,
      "grad_norm": 1.2421875,
      "learning_rate": 5.89419206440483e-06,
      "loss": 0.1043,
      "step": 16340
    },
    {
      "epoch": 4.445350734094617,
      "grad_norm": 1.2890625,
      "learning_rate": 5.8654399079931e-06,
      "loss": 0.0725,
      "step": 16350
    },
    {
      "epoch": 4.445350734094617,
      "eval_loss": 0.11097656190395355,
      "eval_runtime": 191.8758,
      "eval_samples_per_second": 2.606,
      "eval_steps_per_second": 0.083,
      "step": 16350
    },
    {
      "epoch": 4.448069603045133,
      "grad_norm": 1.7890625,
      "learning_rate": 5.836687751581369e-06,
      "loss": 0.1104,
      "step": 16360
    },
    {
      "epoch": 4.45078847199565,
      "grad_norm": 0.93359375,
      "learning_rate": 5.807935595169638e-06,
      "loss": 0.0794,
      "step": 16370
    },
    {
      "epoch": 4.4535073409461665,
      "grad_norm": 1.0546875,
      "learning_rate": 5.779183438757907e-06,
      "loss": 0.0781,
      "step": 16380
    },
    {
      "epoch": 4.456226209896683,
      "grad_norm": 1.5546875,
      "learning_rate": 5.750431282346176e-06,
      "loss": 0.0778,
      "step": 16390
    },
    {
      "epoch": 4.4589450788472,
      "grad_norm": 2.046875,
      "learning_rate": 5.721679125934445e-06,
      "loss": 0.105,
      "step": 16400
    },
    {
      "epoch": 4.4589450788472,
      "eval_loss": 0.11664062738418579,
      "eval_runtime": 105.6103,
      "eval_samples_per_second": 4.734,
      "eval_steps_per_second": 0.152,
      "step": 16400
    },
    {
      "epoch": 4.4616639477977165,
      "grad_norm": 0.6796875,
      "learning_rate": 5.692926969522715e-06,
      "loss": 0.0677,
      "step": 16410
    },
    {
      "epoch": 4.464382816748233,
      "grad_norm": 1.7734375,
      "learning_rate": 5.664174813110983e-06,
      "loss": 0.0602,
      "step": 16420
    },
    {
      "epoch": 4.467101685698749,
      "grad_norm": 1.2890625,
      "learning_rate": 5.635422656699252e-06,
      "loss": 0.0804,
      "step": 16430
    },
    {
      "epoch": 4.4698205546492655,
      "grad_norm": 0.7265625,
      "learning_rate": 5.606670500287522e-06,
      "loss": 0.0705,
      "step": 16440
    },
    {
      "epoch": 4.472539423599782,
      "grad_norm": 1.453125,
      "learning_rate": 5.577918343875791e-06,
      "loss": 0.0967,
      "step": 16450
    },
    {
      "epoch": 4.472539423599782,
      "eval_loss": 0.1247578114271164,
      "eval_runtime": 103.4566,
      "eval_samples_per_second": 4.833,
      "eval_steps_per_second": 0.155,
      "step": 16450
    },
    {
      "epoch": 4.475258292550299,
      "grad_norm": 0.8359375,
      "learning_rate": 5.54916618746406e-06,
      "loss": 0.0886,
      "step": 16460
    },
    {
      "epoch": 4.4779771615008155,
      "grad_norm": 1.734375,
      "learning_rate": 5.520414031052329e-06,
      "loss": 0.0872,
      "step": 16470
    },
    {
      "epoch": 4.480696030451332,
      "grad_norm": 1.1953125,
      "learning_rate": 5.491661874640598e-06,
      "loss": 0.0634,
      "step": 16480
    },
    {
      "epoch": 4.483414899401849,
      "grad_norm": 0.88671875,
      "learning_rate": 5.462909718228867e-06,
      "loss": 0.0748,
      "step": 16490
    },
    {
      "epoch": 4.486133768352365,
      "grad_norm": 0.78515625,
      "learning_rate": 5.434157561817137e-06,
      "loss": 0.071,
      "step": 16500
    },
    {
      "epoch": 4.486133768352365,
      "eval_loss": 0.11760156601667404,
      "eval_runtime": 100.6767,
      "eval_samples_per_second": 4.966,
      "eval_steps_per_second": 0.159,
      "step": 16500
    },
    {
      "epoch": 4.488852637302882,
      "grad_norm": 1.7734375,
      "learning_rate": 5.405405405405406e-06,
      "loss": 0.0699,
      "step": 16510
    },
    {
      "epoch": 4.491571506253399,
      "grad_norm": 2.234375,
      "learning_rate": 5.376653248993674e-06,
      "loss": 0.0912,
      "step": 16520
    },
    {
      "epoch": 4.494290375203915,
      "grad_norm": 1.1484375,
      "learning_rate": 5.347901092581944e-06,
      "loss": 0.0785,
      "step": 16530
    },
    {
      "epoch": 4.497009244154432,
      "grad_norm": 1.6015625,
      "learning_rate": 5.319148936170213e-06,
      "loss": 0.087,
      "step": 16540
    },
    {
      "epoch": 4.499728113104949,
      "grad_norm": 2.078125,
      "learning_rate": 5.290396779758482e-06,
      "loss": 0.0896,
      "step": 16550
    },
    {
      "epoch": 4.499728113104949,
      "eval_loss": 0.12124218791723251,
      "eval_runtime": 242.3564,
      "eval_samples_per_second": 2.063,
      "eval_steps_per_second": 0.066,
      "step": 16550
    },
    {
      "epoch": 4.502446982055465,
      "grad_norm": 2.234375,
      "learning_rate": 5.261644623346752e-06,
      "loss": 0.093,
      "step": 16560
    },
    {
      "epoch": 4.505165851005982,
      "grad_norm": 2.578125,
      "learning_rate": 5.23289246693502e-06,
      "loss": 0.0852,
      "step": 16570
    },
    {
      "epoch": 4.507884719956498,
      "grad_norm": 1.6953125,
      "learning_rate": 5.204140310523289e-06,
      "loss": 0.0809,
      "step": 16580
    },
    {
      "epoch": 4.510603588907014,
      "grad_norm": 1.0625,
      "learning_rate": 5.175388154111559e-06,
      "loss": 0.0596,
      "step": 16590
    },
    {
      "epoch": 4.513322457857531,
      "grad_norm": 1.4375,
      "learning_rate": 5.146635997699828e-06,
      "loss": 0.1029,
      "step": 16600
    },
    {
      "epoch": 4.513322457857531,
      "eval_loss": 0.12617968022823334,
      "eval_runtime": 135.498,
      "eval_samples_per_second": 3.69,
      "eval_steps_per_second": 0.118,
      "step": 16600
    },
    {
      "epoch": 4.516041326808048,
      "grad_norm": 1.5078125,
      "learning_rate": 5.117883841288096e-06,
      "loss": 0.0704,
      "step": 16610
    },
    {
      "epoch": 4.518760195758564,
      "grad_norm": 1.5625,
      "learning_rate": 5.089131684876366e-06,
      "loss": 0.084,
      "step": 16620
    },
    {
      "epoch": 4.521479064709081,
      "grad_norm": 2.875,
      "learning_rate": 5.060379528464635e-06,
      "loss": 0.0931,
      "step": 16630
    },
    {
      "epoch": 4.524197933659598,
      "grad_norm": 1.9921875,
      "learning_rate": 5.031627372052904e-06,
      "loss": 0.0746,
      "step": 16640
    },
    {
      "epoch": 4.526916802610114,
      "grad_norm": 1.6484375,
      "learning_rate": 5.002875215641174e-06,
      "loss": 0.1016,
      "step": 16650
    },
    {
      "epoch": 4.526916802610114,
      "eval_loss": 0.11333593726158142,
      "eval_runtime": 101.9726,
      "eval_samples_per_second": 4.903,
      "eval_steps_per_second": 0.157,
      "step": 16650
    },
    {
      "epoch": 4.529635671560631,
      "grad_norm": 1.3203125,
      "learning_rate": 4.974123059229442e-06,
      "loss": 0.066,
      "step": 16660
    },
    {
      "epoch": 4.532354540511148,
      "grad_norm": 1.5546875,
      "learning_rate": 4.945370902817711e-06,
      "loss": 0.0911,
      "step": 16670
    },
    {
      "epoch": 4.535073409461664,
      "grad_norm": 1.6484375,
      "learning_rate": 4.916618746405981e-06,
      "loss": 0.0696,
      "step": 16680
    },
    {
      "epoch": 4.537792278412181,
      "grad_norm": 0.79296875,
      "learning_rate": 4.88786658999425e-06,
      "loss": 0.0672,
      "step": 16690
    },
    {
      "epoch": 4.540511147362697,
      "grad_norm": 1.3828125,
      "learning_rate": 4.859114433582519e-06,
      "loss": 0.074,
      "step": 16700
    },
    {
      "epoch": 4.540511147362697,
      "eval_loss": 0.12950000166893005,
      "eval_runtime": 157.8956,
      "eval_samples_per_second": 3.167,
      "eval_steps_per_second": 0.101,
      "step": 16700
    },
    {
      "epoch": 4.543230016313213,
      "grad_norm": 2.46875,
      "learning_rate": 4.830362277170788e-06,
      "loss": 0.0857,
      "step": 16710
    },
    {
      "epoch": 4.54594888526373,
      "grad_norm": 1.1328125,
      "learning_rate": 4.801610120759057e-06,
      "loss": 0.0611,
      "step": 16720
    },
    {
      "epoch": 4.548667754214247,
      "grad_norm": 0.66015625,
      "learning_rate": 4.772857964347326e-06,
      "loss": 0.0721,
      "step": 16730
    },
    {
      "epoch": 4.551386623164763,
      "grad_norm": 1.0390625,
      "learning_rate": 4.744105807935596e-06,
      "loss": 0.0852,
      "step": 16740
    },
    {
      "epoch": 4.55410549211528,
      "grad_norm": 2.171875,
      "learning_rate": 4.715353651523865e-06,
      "loss": 0.091,
      "step": 16750
    },
    {
      "epoch": 4.55410549211528,
      "eval_loss": 0.12198437750339508,
      "eval_runtime": 107.2134,
      "eval_samples_per_second": 4.664,
      "eval_steps_per_second": 0.149,
      "step": 16750
    },
    {
      "epoch": 4.556824361065797,
      "grad_norm": 1.4609375,
      "learning_rate": 4.686601495112134e-06,
      "loss": 0.1081,
      "step": 16760
    },
    {
      "epoch": 4.559543230016313,
      "grad_norm": 3.328125,
      "learning_rate": 4.657849338700403e-06,
      "loss": 0.0757,
      "step": 16770
    },
    {
      "epoch": 4.56226209896683,
      "grad_norm": 2.75,
      "learning_rate": 4.629097182288672e-06,
      "loss": 0.0853,
      "step": 16780
    },
    {
      "epoch": 4.5649809679173465,
      "grad_norm": 1.1796875,
      "learning_rate": 4.600345025876941e-06,
      "loss": 0.0934,
      "step": 16790
    },
    {
      "epoch": 4.567699836867863,
      "grad_norm": 3.15625,
      "learning_rate": 4.57159286946521e-06,
      "loss": 0.0963,
      "step": 16800
    },
    {
      "epoch": 4.567699836867863,
      "eval_loss": 0.1070546880364418,
      "eval_runtime": 98.574,
      "eval_samples_per_second": 5.072,
      "eval_steps_per_second": 0.162,
      "step": 16800
    },
    {
      "epoch": 4.57041870581838,
      "grad_norm": 0.54296875,
      "learning_rate": 4.542840713053479e-06,
      "loss": 0.0833,
      "step": 16810
    },
    {
      "epoch": 4.5731375747688965,
      "grad_norm": 2.546875,
      "learning_rate": 4.514088556641749e-06,
      "loss": 0.0933,
      "step": 16820
    },
    {
      "epoch": 4.575856443719413,
      "grad_norm": 2.25,
      "learning_rate": 4.485336400230018e-06,
      "loss": 0.0992,
      "step": 16830
    },
    {
      "epoch": 4.57857531266993,
      "grad_norm": 2.15625,
      "learning_rate": 4.456584243818287e-06,
      "loss": 0.0918,
      "step": 16840
    },
    {
      "epoch": 4.5812941816204455,
      "grad_norm": 1.828125,
      "learning_rate": 4.427832087406556e-06,
      "loss": 0.0879,
      "step": 16850
    },
    {
      "epoch": 4.5812941816204455,
      "eval_loss": 0.11189062148332596,
      "eval_runtime": 133.3472,
      "eval_samples_per_second": 3.75,
      "eval_steps_per_second": 0.12,
      "step": 16850
    },
    {
      "epoch": 4.584013050570962,
      "grad_norm": 2.078125,
      "learning_rate": 4.399079930994825e-06,
      "loss": 0.0731,
      "step": 16860
    },
    {
      "epoch": 4.586731919521479,
      "grad_norm": 0.83203125,
      "learning_rate": 4.370327774583094e-06,
      "loss": 0.0746,
      "step": 16870
    },
    {
      "epoch": 4.5894507884719955,
      "grad_norm": 1.46875,
      "learning_rate": 4.341575618171364e-06,
      "loss": 0.0793,
      "step": 16880
    },
    {
      "epoch": 4.592169657422512,
      "grad_norm": 1.1796875,
      "learning_rate": 4.312823461759633e-06,
      "loss": 0.0818,
      "step": 16890
    },
    {
      "epoch": 4.594888526373029,
      "grad_norm": 1.6640625,
      "learning_rate": 4.284071305347901e-06,
      "loss": 0.0776,
      "step": 16900
    },
    {
      "epoch": 4.594888526373029,
      "eval_loss": 0.12101562321186066,
      "eval_runtime": 185.4533,
      "eval_samples_per_second": 2.696,
      "eval_steps_per_second": 0.086,
      "step": 16900
    },
    {
      "epoch": 4.5976073953235455,
      "grad_norm": 0.8828125,
      "learning_rate": 4.255319148936171e-06,
      "loss": 0.0727,
      "step": 16910
    },
    {
      "epoch": 4.600326264274062,
      "grad_norm": 3.65625,
      "learning_rate": 4.22656699252444e-06,
      "loss": 0.0814,
      "step": 16920
    },
    {
      "epoch": 4.603045133224579,
      "grad_norm": 1.1328125,
      "learning_rate": 4.197814836112709e-06,
      "loss": 0.092,
      "step": 16930
    },
    {
      "epoch": 4.605764002175095,
      "grad_norm": 1.703125,
      "learning_rate": 4.1690626797009785e-06,
      "loss": 0.0758,
      "step": 16940
    },
    {
      "epoch": 4.608482871125612,
      "grad_norm": 1.28125,
      "learning_rate": 4.140310523289247e-06,
      "loss": 0.0847,
      "step": 16950
    },
    {
      "epoch": 4.608482871125612,
      "eval_loss": 0.11503124982118607,
      "eval_runtime": 143.3895,
      "eval_samples_per_second": 3.487,
      "eval_steps_per_second": 0.112,
      "step": 16950
    },
    {
      "epoch": 4.611201740076129,
      "grad_norm": 2.078125,
      "learning_rate": 4.111558366877516e-06,
      "loss": 0.0692,
      "step": 16960
    },
    {
      "epoch": 4.6139206090266445,
      "grad_norm": 3.90625,
      "learning_rate": 4.0828062104657855e-06,
      "loss": 0.0916,
      "step": 16970
    },
    {
      "epoch": 4.616639477977161,
      "grad_norm": 2.03125,
      "learning_rate": 4.0540540540540545e-06,
      "loss": 0.0917,
      "step": 16980
    },
    {
      "epoch": 4.619358346927678,
      "grad_norm": 3.1875,
      "learning_rate": 4.025301897642323e-06,
      "loss": 0.0885,
      "step": 16990
    },
    {
      "epoch": 4.622077215878194,
      "grad_norm": 2.109375,
      "learning_rate": 3.9965497412305925e-06,
      "loss": 0.1029,
      "step": 17000
    },
    {
      "epoch": 4.622077215878194,
      "eval_loss": 0.12467187643051147,
      "eval_runtime": 101.1594,
      "eval_samples_per_second": 4.943,
      "eval_steps_per_second": 0.158,
      "step": 17000
    },
    {
      "epoch": 4.624796084828711,
      "grad_norm": 2.421875,
      "learning_rate": 3.9677975848188615e-06,
      "loss": 0.0924,
      "step": 17010
    },
    {
      "epoch": 4.627514953779228,
      "grad_norm": 2.265625,
      "learning_rate": 3.9390454284071306e-06,
      "loss": 0.1089,
      "step": 17020
    },
    {
      "epoch": 4.630233822729744,
      "grad_norm": 1.390625,
      "learning_rate": 3.9102932719954e-06,
      "loss": 0.0679,
      "step": 17030
    },
    {
      "epoch": 4.632952691680261,
      "grad_norm": 4.0625,
      "learning_rate": 3.8815411155836686e-06,
      "loss": 0.0704,
      "step": 17040
    },
    {
      "epoch": 4.635671560630778,
      "grad_norm": 1.4921875,
      "learning_rate": 3.852788959171938e-06,
      "loss": 0.0754,
      "step": 17050
    },
    {
      "epoch": 4.635671560630778,
      "eval_loss": 0.12296094000339508,
      "eval_runtime": 138.7741,
      "eval_samples_per_second": 3.603,
      "eval_steps_per_second": 0.115,
      "step": 17050
    },
    {
      "epoch": 4.638390429581294,
      "grad_norm": 1.9765625,
      "learning_rate": 3.8240368027602074e-06,
      "loss": 0.0907,
      "step": 17060
    },
    {
      "epoch": 4.641109298531811,
      "grad_norm": 1.953125,
      "learning_rate": 3.7952846463484764e-06,
      "loss": 0.0953,
      "step": 17070
    },
    {
      "epoch": 4.643828167482328,
      "grad_norm": 1.1171875,
      "learning_rate": 3.766532489936746e-06,
      "loss": 0.0914,
      "step": 17080
    },
    {
      "epoch": 4.646547036432844,
      "grad_norm": 2.1875,
      "learning_rate": 3.7377803335250145e-06,
      "loss": 0.0883,
      "step": 17090
    },
    {
      "epoch": 4.649265905383361,
      "grad_norm": 0.369140625,
      "learning_rate": 3.7090281771132835e-06,
      "loss": 0.0749,
      "step": 17100
    },
    {
      "epoch": 4.649265905383361,
      "eval_loss": 0.1109296903014183,
      "eval_runtime": 110.9906,
      "eval_samples_per_second": 4.505,
      "eval_steps_per_second": 0.144,
      "step": 17100
    },
    {
      "epoch": 4.651984774333878,
      "grad_norm": 1.5,
      "learning_rate": 3.680276020701553e-06,
      "loss": 0.089,
      "step": 17110
    },
    {
      "epoch": 4.654703643284393,
      "grad_norm": 1.3046875,
      "learning_rate": 3.651523864289822e-06,
      "loss": 0.094,
      "step": 17120
    },
    {
      "epoch": 4.65742251223491,
      "grad_norm": 1.0859375,
      "learning_rate": 3.6227717078780913e-06,
      "loss": 0.0908,
      "step": 17130
    },
    {
      "epoch": 4.660141381185427,
      "grad_norm": 1.1875,
      "learning_rate": 3.59401955146636e-06,
      "loss": 0.071,
      "step": 17140
    },
    {
      "epoch": 4.662860250135943,
      "grad_norm": 0.64453125,
      "learning_rate": 3.5652673950546294e-06,
      "loss": 0.066,
      "step": 17150
    },
    {
      "epoch": 4.662860250135943,
      "eval_loss": 0.1130390614271164,
      "eval_runtime": 107.2866,
      "eval_samples_per_second": 4.66,
      "eval_steps_per_second": 0.149,
      "step": 17150
    },
    {
      "epoch": 4.66557911908646,
      "grad_norm": 1.671875,
      "learning_rate": 3.5365152386428984e-06,
      "loss": 0.0889,
      "step": 17160
    },
    {
      "epoch": 4.668297988036977,
      "grad_norm": 1.25,
      "learning_rate": 3.507763082231168e-06,
      "loss": 0.0732,
      "step": 17170
    },
    {
      "epoch": 4.671016856987493,
      "grad_norm": 1.75,
      "learning_rate": 3.4790109258194364e-06,
      "loss": 0.0887,
      "step": 17180
    },
    {
      "epoch": 4.67373572593801,
      "grad_norm": 2.796875,
      "learning_rate": 3.4502587694077054e-06,
      "loss": 0.08,
      "step": 17190
    },
    {
      "epoch": 4.6764545948885266,
      "grad_norm": 1.3515625,
      "learning_rate": 3.421506612995975e-06,
      "loss": 0.0732,
      "step": 17200
    },
    {
      "epoch": 4.6764545948885266,
      "eval_loss": 0.11578124761581421,
      "eval_runtime": 107.4992,
      "eval_samples_per_second": 4.651,
      "eval_steps_per_second": 0.149,
      "step": 17200
    },
    {
      "epoch": 4.679173463839043,
      "grad_norm": 1.1328125,
      "learning_rate": 3.3927544565842443e-06,
      "loss": 0.0804,
      "step": 17210
    },
    {
      "epoch": 4.68189233278956,
      "grad_norm": 1.0625,
      "learning_rate": 3.3640023001725133e-06,
      "loss": 0.0773,
      "step": 17220
    },
    {
      "epoch": 4.6846112017400765,
      "grad_norm": 0.61328125,
      "learning_rate": 3.335250143760782e-06,
      "loss": 0.0846,
      "step": 17230
    },
    {
      "epoch": 4.687330070690592,
      "grad_norm": 1.0703125,
      "learning_rate": 3.3064979873490513e-06,
      "loss": 0.0878,
      "step": 17240
    },
    {
      "epoch": 4.690048939641109,
      "grad_norm": 3.03125,
      "learning_rate": 3.2777458309373203e-06,
      "loss": 0.0898,
      "step": 17250
    },
    {
      "epoch": 4.690048939641109,
      "eval_loss": 0.10960156470537186,
      "eval_runtime": 104.9867,
      "eval_samples_per_second": 4.763,
      "eval_steps_per_second": 0.152,
      "step": 17250
    },
    {
      "epoch": 4.692767808591626,
      "grad_norm": 1.1640625,
      "learning_rate": 3.2489936745255897e-06,
      "loss": 0.0823,
      "step": 17260
    },
    {
      "epoch": 4.695486677542142,
      "grad_norm": 1.046875,
      "learning_rate": 3.220241518113859e-06,
      "loss": 0.0769,
      "step": 17270
    },
    {
      "epoch": 4.698205546492659,
      "grad_norm": 1.8671875,
      "learning_rate": 3.1914893617021277e-06,
      "loss": 0.1,
      "step": 17280
    },
    {
      "epoch": 4.7009244154431755,
      "grad_norm": 1.3828125,
      "learning_rate": 3.1627372052903968e-06,
      "loss": 0.0815,
      "step": 17290
    },
    {
      "epoch": 4.703643284393692,
      "grad_norm": 1.6953125,
      "learning_rate": 3.133985048878666e-06,
      "loss": 0.0812,
      "step": 17300
    },
    {
      "epoch": 4.703643284393692,
      "eval_loss": 0.12397656589746475,
      "eval_runtime": 116.8881,
      "eval_samples_per_second": 4.278,
      "eval_steps_per_second": 0.137,
      "step": 17300
    },
    {
      "epoch": 4.706362153344209,
      "grad_norm": 1.5703125,
      "learning_rate": 3.105232892466935e-06,
      "loss": 0.0565,
      "step": 17310
    },
    {
      "epoch": 4.7090810222947255,
      "grad_norm": 1.296875,
      "learning_rate": 3.076480736055204e-06,
      "loss": 0.0995,
      "step": 17320
    },
    {
      "epoch": 4.711799891245242,
      "grad_norm": 1.9609375,
      "learning_rate": 3.0477285796434736e-06,
      "loss": 0.0787,
      "step": 17330
    },
    {
      "epoch": 4.714518760195759,
      "grad_norm": 1.1953125,
      "learning_rate": 3.0189764232317426e-06,
      "loss": 0.0772,
      "step": 17340
    },
    {
      "epoch": 4.717237629146275,
      "grad_norm": 1.375,
      "learning_rate": 2.9902242668200117e-06,
      "loss": 0.1003,
      "step": 17350
    },
    {
      "epoch": 4.717237629146275,
      "eval_loss": 0.11336718499660492,
      "eval_runtime": 129.1106,
      "eval_samples_per_second": 3.873,
      "eval_steps_per_second": 0.124,
      "step": 17350
    },
    {
      "epoch": 4.719956498096792,
      "grad_norm": 1.3359375,
      "learning_rate": 2.9614721104082807e-06,
      "loss": 0.0977,
      "step": 17360
    },
    {
      "epoch": 4.722675367047309,
      "grad_norm": 1.7890625,
      "learning_rate": 2.93271995399655e-06,
      "loss": 0.0732,
      "step": 17370
    },
    {
      "epoch": 4.725394235997825,
      "grad_norm": 1.9375,
      "learning_rate": 2.903967797584819e-06,
      "loss": 0.0672,
      "step": 17380
    },
    {
      "epoch": 4.728113104948341,
      "grad_norm": 2.125,
      "learning_rate": 2.875215641173088e-06,
      "loss": 0.0755,
      "step": 17390
    },
    {
      "epoch": 4.730831973898858,
      "grad_norm": 1.921875,
      "learning_rate": 2.8464634847613575e-06,
      "loss": 0.0885,
      "step": 17400
    },
    {
      "epoch": 4.730831973898858,
      "eval_loss": 0.11213281005620956,
      "eval_runtime": 112.0574,
      "eval_samples_per_second": 4.462,
      "eval_steps_per_second": 0.143,
      "step": 17400
    },
    {
      "epoch": 4.733550842849374,
      "grad_norm": 1.1171875,
      "learning_rate": 2.817711328349626e-06,
      "loss": 0.0599,
      "step": 17410
    },
    {
      "epoch": 4.736269711799891,
      "grad_norm": 1.1328125,
      "learning_rate": 2.7889591719378956e-06,
      "loss": 0.0784,
      "step": 17420
    },
    {
      "epoch": 4.738988580750408,
      "grad_norm": 3.296875,
      "learning_rate": 2.7602070155261646e-06,
      "loss": 0.1004,
      "step": 17430
    },
    {
      "epoch": 4.741707449700924,
      "grad_norm": 1.6484375,
      "learning_rate": 2.7314548591144336e-06,
      "loss": 0.0961,
      "step": 17440
    },
    {
      "epoch": 4.744426318651441,
      "grad_norm": 1.84375,
      "learning_rate": 2.702702702702703e-06,
      "loss": 0.0966,
      "step": 17450
    },
    {
      "epoch": 4.744426318651441,
      "eval_loss": 0.12035156041383743,
      "eval_runtime": 109.0902,
      "eval_samples_per_second": 4.583,
      "eval_steps_per_second": 0.147,
      "step": 17450
    },
    {
      "epoch": 4.747145187601958,
      "grad_norm": 1.578125,
      "learning_rate": 2.673950546290972e-06,
      "loss": 0.0905,
      "step": 17460
    },
    {
      "epoch": 4.749864056552474,
      "grad_norm": 1.8671875,
      "learning_rate": 2.645198389879241e-06,
      "loss": 0.0834,
      "step": 17470
    },
    {
      "epoch": 4.752582925502991,
      "grad_norm": 0.94140625,
      "learning_rate": 2.61644623346751e-06,
      "loss": 0.0636,
      "step": 17480
    },
    {
      "epoch": 4.755301794453508,
      "grad_norm": 1.5625,
      "learning_rate": 2.5876940770557795e-06,
      "loss": 0.077,
      "step": 17490
    },
    {
      "epoch": 4.758020663404024,
      "grad_norm": 1.109375,
      "learning_rate": 2.558941920644048e-06,
      "loss": 0.0672,
      "step": 17500
    },
    {
      "epoch": 4.758020663404024,
      "eval_loss": 0.11603906005620956,
      "eval_runtime": 110.1031,
      "eval_samples_per_second": 4.541,
      "eval_steps_per_second": 0.145,
      "step": 17500
    },
    {
      "epoch": 4.76073953235454,
      "grad_norm": 0.98828125,
      "learning_rate": 2.5301897642323175e-06,
      "loss": 0.0969,
      "step": 17510
    },
    {
      "epoch": 4.763458401305057,
      "grad_norm": 0.498046875,
      "learning_rate": 2.501437607820587e-06,
      "loss": 0.0769,
      "step": 17520
    },
    {
      "epoch": 4.766177270255573,
      "grad_norm": 2.59375,
      "learning_rate": 2.4726854514088555e-06,
      "loss": 0.0858,
      "step": 17530
    },
    {
      "epoch": 4.76889613920609,
      "grad_norm": 2.921875,
      "learning_rate": 2.443933294997125e-06,
      "loss": 0.0791,
      "step": 17540
    },
    {
      "epoch": 4.771615008156607,
      "grad_norm": 2.0,
      "learning_rate": 2.415181138585394e-06,
      "loss": 0.0867,
      "step": 17550
    },
    {
      "epoch": 4.771615008156607,
      "eval_loss": 0.11722656339406967,
      "eval_runtime": 106.122,
      "eval_samples_per_second": 4.712,
      "eval_steps_per_second": 0.151,
      "step": 17550
    },
    {
      "epoch": 4.774333877107123,
      "grad_norm": 2.296875,
      "learning_rate": 2.386428982173663e-06,
      "loss": 0.0972,
      "step": 17560
    },
    {
      "epoch": 4.77705274605764,
      "grad_norm": 2.6875,
      "learning_rate": 2.3576768257619324e-06,
      "loss": 0.0894,
      "step": 17570
    },
    {
      "epoch": 4.779771615008157,
      "grad_norm": 0.9921875,
      "learning_rate": 2.3289246693502014e-06,
      "loss": 0.0922,
      "step": 17580
    },
    {
      "epoch": 4.782490483958673,
      "grad_norm": 2.265625,
      "learning_rate": 2.3001725129384704e-06,
      "loss": 0.0893,
      "step": 17590
    },
    {
      "epoch": 4.78520935290919,
      "grad_norm": 2.328125,
      "learning_rate": 2.2714203565267394e-06,
      "loss": 0.0682,
      "step": 17600
    },
    {
      "epoch": 4.78520935290919,
      "eval_loss": 0.12142187356948853,
      "eval_runtime": 103.3163,
      "eval_samples_per_second": 4.84,
      "eval_steps_per_second": 0.155,
      "step": 17600
    },
    {
      "epoch": 4.787928221859707,
      "grad_norm": 1.8046875,
      "learning_rate": 2.242668200115009e-06,
      "loss": 0.0891,
      "step": 17610
    },
    {
      "epoch": 4.790647090810223,
      "grad_norm": 0.59765625,
      "learning_rate": 2.213916043703278e-06,
      "loss": 0.0663,
      "step": 17620
    },
    {
      "epoch": 4.79336595976074,
      "grad_norm": 1.2578125,
      "learning_rate": 2.185163887291547e-06,
      "loss": 0.094,
      "step": 17630
    },
    {
      "epoch": 4.7960848287112565,
      "grad_norm": 1.03125,
      "learning_rate": 2.1564117308798163e-06,
      "loss": 0.0715,
      "step": 17640
    },
    {
      "epoch": 4.798803697661773,
      "grad_norm": 2.734375,
      "learning_rate": 2.1276595744680853e-06,
      "loss": 0.0969,
      "step": 17650
    },
    {
      "epoch": 4.798803697661773,
      "eval_loss": 0.11992968618869781,
      "eval_runtime": 102.6601,
      "eval_samples_per_second": 4.87,
      "eval_steps_per_second": 0.156,
      "step": 17650
    },
    {
      "epoch": 4.801522566612289,
      "grad_norm": 1.03125,
      "learning_rate": 2.0989074180563543e-06,
      "loss": 0.0728,
      "step": 17660
    },
    {
      "epoch": 4.804241435562806,
      "grad_norm": 1.2109375,
      "learning_rate": 2.0701552616446233e-06,
      "loss": 0.0864,
      "step": 17670
    },
    {
      "epoch": 4.806960304513322,
      "grad_norm": 1.7421875,
      "learning_rate": 2.0414031052328928e-06,
      "loss": 0.0967,
      "step": 17680
    },
    {
      "epoch": 4.809679173463839,
      "grad_norm": 3.765625,
      "learning_rate": 2.0126509488211613e-06,
      "loss": 0.0835,
      "step": 17690
    },
    {
      "epoch": 4.8123980424143555,
      "grad_norm": 1.9140625,
      "learning_rate": 1.9838987924094308e-06,
      "loss": 0.0826,
      "step": 17700
    },
    {
      "epoch": 4.8123980424143555,
      "eval_loss": 0.11042187362909317,
      "eval_runtime": 102.9469,
      "eval_samples_per_second": 4.857,
      "eval_steps_per_second": 0.155,
      "step": 17700
    },
    {
      "epoch": 4.815116911364872,
      "grad_norm": 0.75,
      "learning_rate": 1.9551466359977e-06,
      "loss": 0.0981,
      "step": 17710
    },
    {
      "epoch": 4.817835780315389,
      "grad_norm": 1.2421875,
      "learning_rate": 1.926394479585969e-06,
      "loss": 0.0654,
      "step": 17720
    },
    {
      "epoch": 4.8205546492659055,
      "grad_norm": 1.703125,
      "learning_rate": 1.8976423231742382e-06,
      "loss": 0.0901,
      "step": 17730
    },
    {
      "epoch": 4.823273518216422,
      "grad_norm": 3.046875,
      "learning_rate": 1.8688901667625072e-06,
      "loss": 0.0798,
      "step": 17740
    },
    {
      "epoch": 4.825992387166939,
      "grad_norm": 2.703125,
      "learning_rate": 1.8401380103507765e-06,
      "loss": 0.0923,
      "step": 17750
    },
    {
      "epoch": 4.825992387166939,
      "eval_loss": 0.11085156351327896,
      "eval_runtime": 101.1957,
      "eval_samples_per_second": 4.941,
      "eval_steps_per_second": 0.158,
      "step": 17750
    },
    {
      "epoch": 4.828711256117455,
      "grad_norm": 1.1640625,
      "learning_rate": 1.8113858539390457e-06,
      "loss": 0.0899,
      "step": 17760
    },
    {
      "epoch": 4.831430125067972,
      "grad_norm": 0.95703125,
      "learning_rate": 1.7826336975273147e-06,
      "loss": 0.0837,
      "step": 17770
    },
    {
      "epoch": 4.834148994018488,
      "grad_norm": 0.98828125,
      "learning_rate": 1.753881541115584e-06,
      "loss": 0.0804,
      "step": 17780
    },
    {
      "epoch": 4.8368678629690045,
      "grad_norm": 1.3046875,
      "learning_rate": 1.7251293847038527e-06,
      "loss": 0.0929,
      "step": 17790
    },
    {
      "epoch": 4.839586731919521,
      "grad_norm": 1.125,
      "learning_rate": 1.6963772282921221e-06,
      "loss": 0.0955,
      "step": 17800
    },
    {
      "epoch": 4.839586731919521,
      "eval_loss": 0.11763281375169754,
      "eval_runtime": 101.4114,
      "eval_samples_per_second": 4.93,
      "eval_steps_per_second": 0.158,
      "step": 17800
    },
    {
      "epoch": 4.842305600870038,
      "grad_norm": 0.8125,
      "learning_rate": 1.667625071880391e-06,
      "loss": 0.0908,
      "step": 17810
    },
    {
      "epoch": 4.8450244698205545,
      "grad_norm": 1.34375,
      "learning_rate": 1.6388729154686601e-06,
      "loss": 0.0734,
      "step": 17820
    },
    {
      "epoch": 4.847743338771071,
      "grad_norm": 1.53125,
      "learning_rate": 1.6101207590569296e-06,
      "loss": 0.0956,
      "step": 17830
    },
    {
      "epoch": 4.850462207721588,
      "grad_norm": 1.546875,
      "learning_rate": 1.5813686026451984e-06,
      "loss": 0.102,
      "step": 17840
    },
    {
      "epoch": 4.853181076672104,
      "grad_norm": 1.3046875,
      "learning_rate": 1.5526164462334676e-06,
      "loss": 0.1003,
      "step": 17850
    },
    {
      "epoch": 4.853181076672104,
      "eval_loss": 0.12105468660593033,
      "eval_runtime": 108.7767,
      "eval_samples_per_second": 4.597,
      "eval_steps_per_second": 0.147,
      "step": 17850
    },
    {
      "epoch": 4.855899945622621,
      "grad_norm": 1.3203125,
      "learning_rate": 1.5238642898217368e-06,
      "loss": 0.0604,
      "step": 17860
    },
    {
      "epoch": 4.858618814573138,
      "grad_norm": 0.984375,
      "learning_rate": 1.4951121334100058e-06,
      "loss": 0.0802,
      "step": 17870
    },
    {
      "epoch": 4.861337683523654,
      "grad_norm": 2.25,
      "learning_rate": 1.466359976998275e-06,
      "loss": 0.092,
      "step": 17880
    },
    {
      "epoch": 4.864056552474171,
      "grad_norm": 3.109375,
      "learning_rate": 1.437607820586544e-06,
      "loss": 0.089,
      "step": 17890
    },
    {
      "epoch": 4.866775421424688,
      "grad_norm": 1.46875,
      "learning_rate": 1.408855664174813e-06,
      "loss": 0.0684,
      "step": 17900
    },
    {
      "epoch": 4.866775421424688,
      "eval_loss": 0.1112656220793724,
      "eval_runtime": 109.2038,
      "eval_samples_per_second": 4.579,
      "eval_steps_per_second": 0.147,
      "step": 17900
    },
    {
      "epoch": 4.869494290375204,
      "grad_norm": 1.3125,
      "learning_rate": 1.3801035077630823e-06,
      "loss": 0.0918,
      "step": 17910
    },
    {
      "epoch": 4.872213159325721,
      "grad_norm": 3.15625,
      "learning_rate": 1.3513513513513515e-06,
      "loss": 0.1042,
      "step": 17920
    },
    {
      "epoch": 4.874932028276237,
      "grad_norm": 1.9609375,
      "learning_rate": 1.3225991949396205e-06,
      "loss": 0.1026,
      "step": 17930
    },
    {
      "epoch": 4.877650897226753,
      "grad_norm": 2.125,
      "learning_rate": 1.2938470385278897e-06,
      "loss": 0.0921,
      "step": 17940
    },
    {
      "epoch": 4.88036976617727,
      "grad_norm": 1.890625,
      "learning_rate": 1.2650948821161587e-06,
      "loss": 0.0971,
      "step": 17950
    },
    {
      "epoch": 4.88036976617727,
      "eval_loss": 0.1188671886920929,
      "eval_runtime": 102.9757,
      "eval_samples_per_second": 4.856,
      "eval_steps_per_second": 0.155,
      "step": 17950
    },
    {
      "epoch": 4.883088635127787,
      "grad_norm": 1.8359375,
      "learning_rate": 1.2363427257044278e-06,
      "loss": 0.0905,
      "step": 17960
    },
    {
      "epoch": 4.885807504078303,
      "grad_norm": 0.9609375,
      "learning_rate": 1.207590569292697e-06,
      "loss": 0.0708,
      "step": 17970
    },
    {
      "epoch": 4.88852637302882,
      "grad_norm": 0.98828125,
      "learning_rate": 1.1788384128809662e-06,
      "loss": 0.0947,
      "step": 17980
    },
    {
      "epoch": 4.891245241979337,
      "grad_norm": 1.1953125,
      "learning_rate": 1.1500862564692352e-06,
      "loss": 0.067,
      "step": 17990
    },
    {
      "epoch": 4.893964110929853,
      "grad_norm": 1.2578125,
      "learning_rate": 1.1213341000575044e-06,
      "loss": 0.0787,
      "step": 18000
    },
    {
      "epoch": 4.893964110929853,
      "eval_loss": 0.1107499971985817,
      "eval_runtime": 100.1028,
      "eval_samples_per_second": 4.995,
      "eval_steps_per_second": 0.16,
      "step": 18000
    }
  ],
  "logging_steps": 10,
  "max_steps": 18390,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 8.671136583867552e+18,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}