{
  "best_metric": 0.9323536157608032,
  "best_model_checkpoint": "/scratch/czm5kz/llama2-13b_32_1_0.0003_alternate/checkpoint-1400",
  "epoch": 0.9975062344139651,
  "eval_steps": 20,
  "global_step": 1400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 0.9500238299369812,
      "learning_rate": 0.0002989308624376336,
      "loss": 2.9196,
      "step": 5
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.3018202781677246,
      "learning_rate": 0.00029786172487526725,
      "loss": 2.1494,
      "step": 10
    },
    {
      "epoch": 0.01,
      "grad_norm": 0.9286252856254578,
      "learning_rate": 0.0002967925873129009,
      "loss": 1.7255,
      "step": 15
    },
    {
      "epoch": 0.01,
      "grad_norm": 1.2644277811050415,
      "learning_rate": 0.00029572344975053457,
      "loss": 1.5678,
      "step": 20
    },
    {
      "epoch": 0.01,
      "eval_loss": 1.4308089017868042,
      "eval_runtime": 429.3923,
      "eval_samples_per_second": 26.151,
      "eval_steps_per_second": 3.27,
      "step": 20
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.486525535583496,
      "learning_rate": 0.00029465431218816815,
      "loss": 1.3281,
      "step": 25
    },
    {
      "epoch": 0.02,
      "grad_norm": 1.1946622133255005,
      "learning_rate": 0.00029358517462580184,
      "loss": 1.2098,
      "step": 30
    },
    {
      "epoch": 0.02,
      "grad_norm": 0.7685691118240356,
      "learning_rate": 0.0002925160370634355,
      "loss": 1.2045,
      "step": 35
    },
    {
      "epoch": 0.03,
      "grad_norm": 1.084057092666626,
      "learning_rate": 0.0002914468995010691,
      "loss": 1.2187,
      "step": 40
    },
    {
      "epoch": 0.03,
      "eval_loss": 1.1283082962036133,
      "eval_runtime": 429.1606,
      "eval_samples_per_second": 26.165,
      "eval_steps_per_second": 3.272,
      "step": 40
    },
    {
      "epoch": 0.03,
      "grad_norm": 0.6618586778640747,
      "learning_rate": 0.00029037776193870275,
      "loss": 1.1625,
      "step": 45
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.5745049118995667,
      "learning_rate": 0.0002893086243763364,
      "loss": 1.1622,
      "step": 50
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.6866307854652405,
      "learning_rate": 0.00028823948681397,
      "loss": 1.1375,
      "step": 55
    },
    {
      "epoch": 0.04,
      "grad_norm": 0.7473587393760681,
      "learning_rate": 0.0002871703492516037,
      "loss": 1.1048,
      "step": 60
    },
    {
      "epoch": 0.04,
      "eval_loss": 1.0678260326385498,
      "eval_runtime": 430.0891,
      "eval_samples_per_second": 26.109,
      "eval_steps_per_second": 3.264,
      "step": 60
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.5050413012504578,
      "learning_rate": 0.0002861012116892373,
      "loss": 1.082,
      "step": 65
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.546146810054779,
      "learning_rate": 0.000285032074126871,
      "loss": 1.0399,
      "step": 70
    },
    {
      "epoch": 0.05,
      "grad_norm": 0.6042059659957886,
      "learning_rate": 0.0002839629365645046,
      "loss": 0.9998,
      "step": 75
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.5676208734512329,
      "learning_rate": 0.00028289379900213826,
      "loss": 1.0594,
      "step": 80
    },
    {
      "epoch": 0.06,
      "eval_loss": 1.0362563133239746,
      "eval_runtime": 429.9585,
      "eval_samples_per_second": 26.116,
      "eval_steps_per_second": 3.265,
      "step": 80
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.6432541012763977,
      "learning_rate": 0.0002818246614397719,
      "loss": 1.052,
      "step": 85
    },
    {
      "epoch": 0.06,
      "grad_norm": 0.5946438312530518,
      "learning_rate": 0.00028075552387740553,
      "loss": 1.038,
      "step": 90
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.6687917113304138,
      "learning_rate": 0.00027968638631503917,
      "loss": 1.0949,
      "step": 95
    },
    {
      "epoch": 0.07,
      "grad_norm": 1.521995186805725,
      "learning_rate": 0.0002786172487526728,
      "loss": 1.0393,
      "step": 100
    },
    {
      "epoch": 0.07,
      "eval_loss": 1.0224562883377075,
      "eval_runtime": 430.7099,
      "eval_samples_per_second": 26.071,
      "eval_steps_per_second": 3.26,
      "step": 100
    },
    {
      "epoch": 0.07,
      "grad_norm": 0.5327974557876587,
      "learning_rate": 0.00027754811119030644,
      "loss": 1.0709,
      "step": 105
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.47005364298820496,
      "learning_rate": 0.0002764789736279401,
      "loss": 1.0119,
      "step": 110
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.6974779963493347,
      "learning_rate": 0.00027540983606557377,
      "loss": 1.0285,
      "step": 115
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.568474531173706,
      "learning_rate": 0.0002743406985032074,
      "loss": 1.085,
      "step": 120
    },
    {
      "epoch": 0.09,
      "eval_loss": 1.012302041053772,
      "eval_runtime": 430.9145,
      "eval_samples_per_second": 26.059,
      "eval_steps_per_second": 3.258,
      "step": 120
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.4923990070819855,
      "learning_rate": 0.00027327156094084104,
      "loss": 1.0179,
      "step": 125
    },
    {
      "epoch": 0.09,
      "grad_norm": 0.4918578565120697,
      "learning_rate": 0.0002722024233784747,
      "loss": 0.982,
      "step": 130
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.6147764921188354,
      "learning_rate": 0.0002711332858161083,
      "loss": 1.0864,
      "step": 135
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.5552188158035278,
      "learning_rate": 0.00027006414825374195,
      "loss": 1.0125,
      "step": 140
    },
    {
      "epoch": 0.1,
      "eval_loss": 1.004356861114502,
      "eval_runtime": 428.4385,
      "eval_samples_per_second": 26.209,
      "eval_steps_per_second": 3.277,
      "step": 140
    },
    {
      "epoch": 0.1,
      "grad_norm": 0.5992436408996582,
      "learning_rate": 0.00026899501069137564,
      "loss": 1.0427,
      "step": 145
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.44927549362182617,
      "learning_rate": 0.0002679258731290092,
      "loss": 0.9857,
      "step": 150
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.7258988618850708,
      "learning_rate": 0.0002668567355666429,
      "loss": 1.0241,
      "step": 155
    },
    {
      "epoch": 0.11,
      "grad_norm": 0.4783088266849518,
      "learning_rate": 0.00026578759800427654,
      "loss": 0.9954,
      "step": 160
    },
    {
      "epoch": 0.11,
      "eval_loss": 0.9951605796813965,
      "eval_runtime": 429.3957,
      "eval_samples_per_second": 26.151,
      "eval_steps_per_second": 3.27,
      "step": 160
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.547055184841156,
      "learning_rate": 0.0002647184604419102,
      "loss": 1.008,
      "step": 165
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.5691074132919312,
      "learning_rate": 0.0002636493228795438,
      "loss": 0.9818,
      "step": 170
    },
    {
      "epoch": 0.12,
      "grad_norm": 0.5176872611045837,
      "learning_rate": 0.00026258018531717745,
      "loss": 1.009,
      "step": 175
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.469517320394516,
      "learning_rate": 0.0002615110477548111,
      "loss": 1.0347,
      "step": 180
    },
    {
      "epoch": 0.13,
      "eval_loss": 0.9926482439041138,
      "eval_runtime": 428.9611,
      "eval_samples_per_second": 26.177,
      "eval_steps_per_second": 3.273,
      "step": 180
    },
    {
      "epoch": 0.13,
      "grad_norm": 0.4130084812641144,
      "learning_rate": 0.0002604419101924447,
      "loss": 0.9957,
      "step": 185
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.43157169222831726,
      "learning_rate": 0.00025937277263007836,
      "loss": 1.0193,
      "step": 190
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.4014897644519806,
      "learning_rate": 0.000258303635067712,
      "loss": 1.0301,
      "step": 195
    },
    {
      "epoch": 0.14,
      "grad_norm": 0.4352978765964508,
      "learning_rate": 0.0002572344975053457,
      "loss": 0.9812,
      "step": 200
    },
    {
      "epoch": 0.14,
      "eval_loss": 0.9875041246414185,
      "eval_runtime": 429.7836,
      "eval_samples_per_second": 26.127,
      "eval_steps_per_second": 3.267,
      "step": 200
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.350909948348999,
      "learning_rate": 0.0002561653599429793,
      "loss": 0.9866,
      "step": 205
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.48860958218574524,
      "learning_rate": 0.00025509622238061296,
      "loss": 0.9854,
      "step": 210
    },
    {
      "epoch": 0.15,
      "grad_norm": 0.5793005228042603,
      "learning_rate": 0.0002540270848182466,
      "loss": 1.0233,
      "step": 215
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.4061422646045685,
      "learning_rate": 0.00025295794725588023,
      "loss": 0.9912,
      "step": 220
    },
    {
      "epoch": 0.16,
      "eval_loss": 0.9817266464233398,
      "eval_runtime": 428.9163,
      "eval_samples_per_second": 26.18,
      "eval_steps_per_second": 3.273,
      "step": 220
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.41655540466308594,
      "learning_rate": 0.00025188880969351387,
      "loss": 0.9682,
      "step": 225
    },
    {
      "epoch": 0.16,
      "grad_norm": 0.40418338775634766,
      "learning_rate": 0.00025081967213114756,
      "loss": 0.9841,
      "step": 230
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.4219729006290436,
      "learning_rate": 0.00024975053456878114,
      "loss": 0.9733,
      "step": 235
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.4193793833255768,
      "learning_rate": 0.00024868139700641483,
      "loss": 1.0167,
      "step": 240
    },
    {
      "epoch": 0.17,
      "eval_loss": 0.9779065847396851,
      "eval_runtime": 429.6736,
      "eval_samples_per_second": 26.134,
      "eval_steps_per_second": 3.268,
      "step": 240
    },
    {
      "epoch": 0.17,
      "grad_norm": 0.38435232639312744,
      "learning_rate": 0.00024761225944404847,
      "loss": 0.9686,
      "step": 245
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.36323386430740356,
      "learning_rate": 0.0002465431218816821,
      "loss": 0.9785,
      "step": 250
    },
    {
      "epoch": 0.18,
      "grad_norm": 0.4036902189254761,
      "learning_rate": 0.00024547398431931574,
      "loss": 0.9708,
      "step": 255
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.46558523178100586,
      "learning_rate": 0.0002444048467569494,
      "loss": 0.9634,
      "step": 260
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.9745126962661743,
      "eval_runtime": 429.4898,
      "eval_samples_per_second": 26.145,
      "eval_steps_per_second": 3.269,
      "step": 260
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.5028152465820312,
      "learning_rate": 0.000243335709194583,
      "loss": 0.9661,
      "step": 265
    },
    {
      "epoch": 0.19,
      "grad_norm": 0.49994298815727234,
      "learning_rate": 0.00024226657163221665,
      "loss": 1.0082,
      "step": 270
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.41407310962677,
      "learning_rate": 0.0002411974340698503,
      "loss": 0.9414,
      "step": 275
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.39765608310699463,
      "learning_rate": 0.00024012829650748392,
      "loss": 0.9785,
      "step": 280
    },
    {
      "epoch": 0.2,
      "eval_loss": 0.9723387956619263,
      "eval_runtime": 429.7271,
      "eval_samples_per_second": 26.131,
      "eval_steps_per_second": 3.267,
      "step": 280
    },
    {
      "epoch": 0.2,
      "grad_norm": 0.36988940834999084,
      "learning_rate": 0.00023905915894511758,
      "loss": 0.9391,
      "step": 285
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.4206317365169525,
      "learning_rate": 0.00023799002138275122,
      "loss": 0.9463,
      "step": 290
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.45045849680900574,
      "learning_rate": 0.00023692088382038488,
      "loss": 0.9946,
      "step": 295
    },
    {
      "epoch": 0.21,
      "grad_norm": 0.39236941933631897,
      "learning_rate": 0.0002358517462580185,
      "loss": 0.976,
      "step": 300
    },
    {
      "epoch": 0.21,
      "eval_loss": 0.9704678654670715,
      "eval_runtime": 429.6884,
      "eval_samples_per_second": 26.133,
      "eval_steps_per_second": 3.267,
      "step": 300
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.5147037506103516,
      "learning_rate": 0.00023478260869565215,
      "loss": 0.9661,
      "step": 305
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.4205746352672577,
      "learning_rate": 0.0002337134711332858,
      "loss": 1.0082,
      "step": 310
    },
    {
      "epoch": 0.22,
      "grad_norm": 0.5038707852363586,
      "learning_rate": 0.00023264433357091945,
      "loss": 1.0248,
      "step": 315
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.38720741868019104,
      "learning_rate": 0.0002315751960085531,
      "loss": 0.9593,
      "step": 320
    },
    {
      "epoch": 0.23,
      "eval_loss": 0.9690603017807007,
      "eval_runtime": 429.2915,
      "eval_samples_per_second": 26.157,
      "eval_steps_per_second": 3.271,
      "step": 320
    },
    {
      "epoch": 0.23,
      "grad_norm": 0.37954407930374146,
      "learning_rate": 0.00023050605844618672,
      "loss": 0.9919,
      "step": 325
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.9392601251602173,
      "learning_rate": 0.00022943692088382036,
      "loss": 1.003,
      "step": 330
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.5109981894493103,
      "learning_rate": 0.00022836778332145402,
      "loss": 0.9723,
      "step": 335
    },
    {
      "epoch": 0.24,
      "grad_norm": 0.56329345703125,
      "learning_rate": 0.00022729864575908766,
      "loss": 1.0046,
      "step": 340
    },
    {
      "epoch": 0.24,
      "eval_loss": 0.9681074619293213,
      "eval_runtime": 429.1046,
      "eval_samples_per_second": 26.168,
      "eval_steps_per_second": 3.272,
      "step": 340
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.3970394730567932,
      "learning_rate": 0.00022622950819672127,
      "loss": 1.0459,
      "step": 345
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.33091840147972107,
      "learning_rate": 0.00022516037063435493,
      "loss": 0.959,
      "step": 350
    },
    {
      "epoch": 0.25,
      "grad_norm": 0.39255568385124207,
      "learning_rate": 0.00022409123307198857,
      "loss": 0.9297,
      "step": 355
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.40491119027137756,
      "learning_rate": 0.00022302209550962223,
      "loss": 0.9573,
      "step": 360
    },
    {
      "epoch": 0.26,
      "eval_loss": 0.964959442615509,
      "eval_runtime": 429.8883,
      "eval_samples_per_second": 26.121,
      "eval_steps_per_second": 3.266,
      "step": 360
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.3292306065559387,
      "learning_rate": 0.00022195295794725584,
      "loss": 0.9592,
      "step": 365
    },
    {
      "epoch": 0.26,
      "grad_norm": 0.38449323177337646,
      "learning_rate": 0.0002208838203848895,
      "loss": 0.9375,
      "step": 370
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.4422217309474945,
      "learning_rate": 0.00021981468282252314,
      "loss": 0.9794,
      "step": 375
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.5002702474594116,
      "learning_rate": 0.0002187455452601568,
      "loss": 0.9956,
      "step": 380
    },
    {
      "epoch": 0.27,
      "eval_loss": 0.9628071188926697,
      "eval_runtime": 429.2302,
      "eval_samples_per_second": 26.161,
      "eval_steps_per_second": 3.271,
      "step": 380
    },
    {
      "epoch": 0.27,
      "grad_norm": 0.32773861289024353,
      "learning_rate": 0.0002176764076977904,
      "loss": 0.9852,
      "step": 385
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.5032110214233398,
      "learning_rate": 0.00021660727013542407,
      "loss": 0.9525,
      "step": 390
    },
    {
      "epoch": 0.28,
      "grad_norm": 0.3590690791606903,
      "learning_rate": 0.0002155381325730577,
      "loss": 0.9591,
      "step": 395
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.7039294242858887,
      "learning_rate": 0.00021446899501069137,
      "loss": 0.963,
      "step": 400
    },
    {
      "epoch": 0.29,
      "eval_loss": 0.9601727724075317,
      "eval_runtime": 429.5145,
      "eval_samples_per_second": 26.143,
      "eval_steps_per_second": 3.269,
      "step": 400
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.36725956201553345,
      "learning_rate": 0.00021339985744832498,
      "loss": 0.9859,
      "step": 405
    },
    {
      "epoch": 0.29,
      "grad_norm": 0.323743611574173,
      "learning_rate": 0.00021233071988595865,
      "loss": 0.9349,
      "step": 410
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.46232929825782776,
      "learning_rate": 0.00021126158232359228,
      "loss": 0.9344,
      "step": 415
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.408657044172287,
      "learning_rate": 0.00021019244476122595,
      "loss": 0.9849,
      "step": 420
    },
    {
      "epoch": 0.3,
      "eval_loss": 0.9595761299133301,
      "eval_runtime": 428.7767,
      "eval_samples_per_second": 26.188,
      "eval_steps_per_second": 3.274,
      "step": 420
    },
    {
      "epoch": 0.3,
      "grad_norm": 0.3261207640171051,
      "learning_rate": 0.00020912330719885958,
      "loss": 0.9458,
      "step": 425
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.38089486956596375,
      "learning_rate": 0.0002080541696364932,
      "loss": 0.9789,
      "step": 430
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.33345696330070496,
      "learning_rate": 0.00020698503207412685,
      "loss": 0.9492,
      "step": 435
    },
    {
      "epoch": 0.31,
      "grad_norm": 0.4079163074493408,
      "learning_rate": 0.0002059158945117605,
      "loss": 0.9628,
      "step": 440
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.9577447175979614,
      "eval_runtime": 428.7333,
      "eval_samples_per_second": 26.191,
      "eval_steps_per_second": 3.275,
      "step": 440
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.3896418511867523,
      "learning_rate": 0.00020484675694939415,
      "loss": 0.9968,
      "step": 445
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.48451879620552063,
      "learning_rate": 0.00020377761938702776,
      "loss": 0.9276,
      "step": 450
    },
    {
      "epoch": 0.32,
      "grad_norm": 0.3475559651851654,
      "learning_rate": 0.00020270848182466143,
      "loss": 0.9545,
      "step": 455
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.39289623498916626,
      "learning_rate": 0.00020163934426229506,
      "loss": 0.942,
      "step": 460
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.9580677151679993,
      "eval_runtime": 429.8733,
      "eval_samples_per_second": 26.122,
      "eval_steps_per_second": 3.266,
      "step": 460
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.4262286424636841,
      "learning_rate": 0.00020057020669992872,
      "loss": 0.9626,
      "step": 465
    },
    {
      "epoch": 0.33,
      "grad_norm": 0.32213443517684937,
      "learning_rate": 0.00019950106913756233,
      "loss": 0.9654,
      "step": 470
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.5463011264801025,
      "learning_rate": 0.000198431931575196,
      "loss": 0.9739,
      "step": 475
    },
    {
      "epoch": 0.34,
      "grad_norm": 0.43018388748168945,
      "learning_rate": 0.00019736279401282963,
      "loss": 0.9712,
      "step": 480
    },
    {
      "epoch": 0.34,
      "eval_loss": 0.955605149269104,
      "eval_runtime": 430.0606,
      "eval_samples_per_second": 26.11,
      "eval_steps_per_second": 3.265,
      "step": 480
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.3303755819797516,
      "learning_rate": 0.0001962936564504633,
      "loss": 0.9799,
      "step": 485
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.35282695293426514,
      "learning_rate": 0.0001952245188880969,
      "loss": 0.9797,
      "step": 490
    },
    {
      "epoch": 0.35,
      "grad_norm": 0.7973325848579407,
      "learning_rate": 0.00019415538132573057,
      "loss": 0.9733,
      "step": 495
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.46459755301475525,
      "learning_rate": 0.0001930862437633642,
      "loss": 0.9368,
      "step": 500
    },
    {
      "epoch": 0.36,
      "eval_loss": 0.9580536484718323,
      "eval_runtime": 429.9689,
      "eval_samples_per_second": 26.116,
      "eval_steps_per_second": 3.265,
      "step": 500
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.4230199158191681,
      "learning_rate": 0.00019201710620099787,
      "loss": 0.9822,
      "step": 505
    },
    {
      "epoch": 0.36,
      "grad_norm": 0.3397184908390045,
      "learning_rate": 0.00019094796863863148,
      "loss": 0.9522,
      "step": 510
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.6563906073570251,
      "learning_rate": 0.0001898788310762651,
      "loss": 1.0053,
      "step": 515
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.29375872015953064,
      "learning_rate": 0.00018880969351389878,
      "loss": 0.9623,
      "step": 520
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.9542292356491089,
      "eval_runtime": 429.8107,
      "eval_samples_per_second": 26.125,
      "eval_steps_per_second": 3.267,
      "step": 520
    },
    {
      "epoch": 0.37,
      "grad_norm": 0.37316498160362244,
      "learning_rate": 0.0001877405559515324,
      "loss": 0.9615,
      "step": 525
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.447251558303833,
      "learning_rate": 0.00018667141838916605,
      "loss": 0.9654,
      "step": 530
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.4220346510410309,
      "learning_rate": 0.00018560228082679968,
      "loss": 0.9912,
      "step": 535
    },
    {
      "epoch": 0.38,
      "grad_norm": 0.4119402766227722,
      "learning_rate": 0.00018453314326443335,
      "loss": 0.9884,
      "step": 540
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.9527921080589294,
      "eval_runtime": 429.6973,
      "eval_samples_per_second": 26.132,
      "eval_steps_per_second": 3.267,
      "step": 540
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.370868057012558,
      "learning_rate": 0.00018346400570206698,
      "loss": 0.9405,
      "step": 545
    },
    {
      "epoch": 0.39,
      "grad_norm": 0.3117481768131256,
      "learning_rate": 0.00018239486813970065,
      "loss": 0.951,
      "step": 550
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.3839544355869293,
      "learning_rate": 0.00018132573057733425,
      "loss": 0.9802,
      "step": 555
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.28429457545280457,
      "learning_rate": 0.00018025659301496792,
      "loss": 0.9313,
      "step": 560
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.9514893293380737,
      "eval_runtime": 429.9336,
      "eval_samples_per_second": 26.118,
      "eval_steps_per_second": 3.266,
      "step": 560
    },
    {
      "epoch": 0.4,
      "grad_norm": 0.36404356360435486,
      "learning_rate": 0.00017918745545260155,
      "loss": 0.9582,
      "step": 565
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.33048421144485474,
      "learning_rate": 0.00017811831789023522,
      "loss": 0.95,
      "step": 570
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.3449236750602722,
      "learning_rate": 0.00017704918032786883,
      "loss": 0.9765,
      "step": 575
    },
    {
      "epoch": 0.41,
      "grad_norm": 0.3965567350387573,
      "learning_rate": 0.0001759800427655025,
      "loss": 0.9743,
      "step": 580
    },
    {
      "epoch": 0.41,
      "eval_loss": 0.9508348107337952,
      "eval_runtime": 429.4982,
      "eval_samples_per_second": 26.144,
      "eval_steps_per_second": 3.269,
      "step": 580
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.491098552942276,
      "learning_rate": 0.00017491090520313613,
      "loss": 0.9843,
      "step": 585
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.3687509298324585,
      "learning_rate": 0.0001738417676407698,
      "loss": 0.9936,
      "step": 590
    },
    {
      "epoch": 0.42,
      "grad_norm": 0.33909574151039124,
      "learning_rate": 0.0001727726300784034,
      "loss": 0.9334,
      "step": 595
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.36044976115226746,
      "learning_rate": 0.00017170349251603703,
      "loss": 0.9805,
      "step": 600
    },
    {
      "epoch": 0.43,
      "eval_loss": 0.9498609900474548,
      "eval_runtime": 429.1598,
      "eval_samples_per_second": 26.165,
      "eval_steps_per_second": 3.272,
      "step": 600
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.28610169887542725,
      "learning_rate": 0.0001706343549536707,
      "loss": 0.9513,
      "step": 605
    },
    {
      "epoch": 0.43,
      "grad_norm": 0.29482224583625793,
      "learning_rate": 0.00016956521739130433,
      "loss": 0.9313,
      "step": 610
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.46972745656967163,
      "learning_rate": 0.00016849607982893797,
      "loss": 0.9753,
      "step": 615
    },
    {
      "epoch": 0.44,
      "grad_norm": 0.38324683904647827,
      "learning_rate": 0.0001674269422665716,
      "loss": 0.9759,
      "step": 620
    },
    {
      "epoch": 0.44,
      "eval_loss": 0.949146568775177,
      "eval_runtime": 430.2097,
      "eval_samples_per_second": 26.101,
      "eval_steps_per_second": 3.264,
      "step": 620
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.4306158125400543,
      "learning_rate": 0.00016635780470420527,
      "loss": 0.9788,
      "step": 625
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.49315810203552246,
      "learning_rate": 0.0001652886671418389,
      "loss": 0.9636,
      "step": 630
    },
    {
      "epoch": 0.45,
      "grad_norm": 0.4102541506290436,
      "learning_rate": 0.00016421952957947254,
      "loss": 0.9957,
      "step": 635
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.40014615654945374,
      "learning_rate": 0.00016315039201710618,
      "loss": 0.9468,
      "step": 640
    },
    {
      "epoch": 0.46,
      "eval_loss": 0.9478526711463928,
      "eval_runtime": 429.5532,
      "eval_samples_per_second": 26.141,
      "eval_steps_per_second": 3.269,
      "step": 640
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.3694494962692261,
      "learning_rate": 0.00016208125445473984,
      "loss": 0.9824,
      "step": 645
    },
    {
      "epoch": 0.46,
      "grad_norm": 0.37701812386512756,
      "learning_rate": 0.00016101211689237348,
      "loss": 0.9274,
      "step": 650
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.3787167966365814,
      "learning_rate": 0.00015994297933000714,
      "loss": 0.9719,
      "step": 655
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.48614123463630676,
      "learning_rate": 0.00015887384176764075,
      "loss": 0.9593,
      "step": 660
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.9467563629150391,
      "eval_runtime": 430.6789,
      "eval_samples_per_second": 26.073,
      "eval_steps_per_second": 3.26,
      "step": 660
    },
    {
      "epoch": 0.47,
      "grad_norm": 0.35638296604156494,
      "learning_rate": 0.0001578047042052744,
      "loss": 0.9862,
      "step": 665
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.27917659282684326,
      "learning_rate": 0.00015673556664290805,
      "loss": 0.9504,
      "step": 670
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.34654778242111206,
      "learning_rate": 0.0001556664290805417,
      "loss": 0.9428,
      "step": 675
    },
    {
      "epoch": 0.48,
      "grad_norm": 0.2783154845237732,
      "learning_rate": 0.00015459729151817532,
      "loss": 0.9242,
      "step": 680
    },
    {
      "epoch": 0.48,
      "eval_loss": 0.9472882151603699,
      "eval_runtime": 428.9437,
      "eval_samples_per_second": 26.178,
      "eval_steps_per_second": 3.273,
      "step": 680
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.3296685814857483,
      "learning_rate": 0.00015352815395580896,
      "loss": 0.9817,
      "step": 685
    },
    {
      "epoch": 0.49,
      "grad_norm": 0.35922589898109436,
      "learning_rate": 0.00015245901639344262,
      "loss": 0.904,
      "step": 690
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.33108726143836975,
      "learning_rate": 0.00015138987883107623,
      "loss": 0.9537,
      "step": 695
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.3146369457244873,
      "learning_rate": 0.0001503207412687099,
      "loss": 0.9594,
      "step": 700
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.9462445974349976,
      "eval_runtime": 430.3498,
      "eval_samples_per_second": 26.093,
      "eval_steps_per_second": 3.262,
      "step": 700
    },
    {
      "epoch": 0.5,
      "grad_norm": 0.3020135164260864,
      "learning_rate": 0.00014925160370634355,
      "loss": 0.9947,
      "step": 705
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.33066317439079285,
      "learning_rate": 0.0001481824661439772,
      "loss": 0.926,
      "step": 710
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.3479137420654297,
      "learning_rate": 0.00014711332858161083,
      "loss": 0.9388,
      "step": 715
    },
    {
      "epoch": 0.51,
      "grad_norm": 0.4258211851119995,
      "learning_rate": 0.00014604419101924446,
      "loss": 0.9587,
      "step": 720
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.9459365606307983,
      "eval_runtime": 429.6711,
      "eval_samples_per_second": 26.134,
      "eval_steps_per_second": 3.268,
      "step": 720
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.33177533745765686,
      "learning_rate": 0.0001449750534568781,
      "loss": 0.9466,
      "step": 725
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.36505717039108276,
      "learning_rate": 0.00014390591589451173,
      "loss": 0.9591,
      "step": 730
    },
    {
      "epoch": 0.52,
      "grad_norm": 0.2933894693851471,
      "learning_rate": 0.0001428367783321454,
      "loss": 0.9857,
      "step": 735
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.30695998668670654,
      "learning_rate": 0.00014176764076977903,
      "loss": 0.9311,
      "step": 740
    },
    {
      "epoch": 0.53,
      "eval_loss": 0.9447991251945496,
      "eval_runtime": 429.4504,
      "eval_samples_per_second": 26.147,
      "eval_steps_per_second": 3.269,
      "step": 740
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.37066712975502014,
      "learning_rate": 0.00014069850320741267,
      "loss": 0.9075,
      "step": 745
    },
    {
      "epoch": 0.53,
      "grad_norm": 0.3553217053413391,
      "learning_rate": 0.0001396293656450463,
      "loss": 0.9355,
      "step": 750
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.3907565176486969,
      "learning_rate": 0.00013856022808267997,
      "loss": 0.9593,
      "step": 755
    },
    {
      "epoch": 0.54,
      "grad_norm": 0.35674384236335754,
      "learning_rate": 0.0001374910905203136,
      "loss": 0.9485,
      "step": 760
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.9437533617019653,
      "eval_runtime": 430.5708,
      "eval_samples_per_second": 26.079,
      "eval_steps_per_second": 3.261,
      "step": 760
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.39057737588882446,
      "learning_rate": 0.00013642195295794724,
      "loss": 0.9201,
      "step": 765
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.2841975688934326,
      "learning_rate": 0.00013535281539558088,
      "loss": 0.9466,
      "step": 770
    },
    {
      "epoch": 0.55,
      "grad_norm": 0.38508614897727966,
      "learning_rate": 0.00013428367783321454,
      "loss": 0.9772,
      "step": 775
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.3465386629104614,
      "learning_rate": 0.00013321454027084818,
      "loss": 0.9974,
      "step": 780
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.9442404508590698,
      "eval_runtime": 429.5011,
      "eval_samples_per_second": 26.144,
      "eval_steps_per_second": 3.269,
      "step": 780
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.360097199678421,
      "learning_rate": 0.0001321454027084818,
      "loss": 0.9725,
      "step": 785
    },
    {
      "epoch": 0.56,
      "grad_norm": 0.34066542983055115,
      "learning_rate": 0.00013107626514611545,
      "loss": 1.0158,
      "step": 790
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.3403652608394623,
      "learning_rate": 0.0001300071275837491,
      "loss": 0.927,
      "step": 795
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.33237767219543457,
      "learning_rate": 0.00012893799002138275,
      "loss": 0.9516,
      "step": 800
    },
    {
      "epoch": 0.57,
      "eval_loss": 0.9429513216018677,
      "eval_runtime": 429.3919,
      "eval_samples_per_second": 26.151,
      "eval_steps_per_second": 3.27,
      "step": 800
    },
    {
      "epoch": 0.57,
      "grad_norm": 0.33179551362991333,
      "learning_rate": 0.00012786885245901638,
      "loss": 0.922,
      "step": 805
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.3568633794784546,
      "learning_rate": 0.00012679971489665002,
      "loss": 0.9434,
      "step": 810
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.4287993907928467,
      "learning_rate": 0.00012573057733428366,
      "loss": 0.9496,
      "step": 815
    },
    {
      "epoch": 0.58,
      "grad_norm": 0.3873540461063385,
      "learning_rate": 0.0001246614397719173,
      "loss": 0.9476,
      "step": 820
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.942284107208252,
      "eval_runtime": 429.6801,
      "eval_samples_per_second": 26.133,
      "eval_steps_per_second": 3.268,
      "step": 820
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.5366536974906921,
      "learning_rate": 0.00012359230220955095,
      "loss": 0.9609,
      "step": 825
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.38116729259490967,
      "learning_rate": 0.0001225231646471846,
      "loss": 0.946,
      "step": 830
    },
    {
      "epoch": 0.59,
      "grad_norm": 0.3352102041244507,
      "learning_rate": 0.00012145402708481824,
      "loss": 0.9487,
      "step": 835
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.314010888338089,
      "learning_rate": 0.00012038488952245188,
      "loss": 0.951,
      "step": 840
    },
    {
      "epoch": 0.6,
      "eval_loss": 0.9439210295677185,
      "eval_runtime": 429.6236,
      "eval_samples_per_second": 26.137,
      "eval_steps_per_second": 3.268,
      "step": 840
    },
    {
      "epoch": 0.6,
      "grad_norm": 0.359325647354126,
      "learning_rate": 0.00011931575196008553,
      "loss": 0.9628,
      "step": 845
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.5471125841140747,
      "learning_rate": 0.00011824661439771916,
      "loss": 0.9581,
      "step": 850
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.5925132632255554,
      "learning_rate": 0.00011717747683535281,
      "loss": 0.9777,
      "step": 855
    },
    {
      "epoch": 0.61,
      "grad_norm": 0.3140357732772827,
      "learning_rate": 0.00011610833927298645,
      "loss": 0.9269,
      "step": 860
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.9412524700164795,
      "eval_runtime": 429.1443,
      "eval_samples_per_second": 26.166,
      "eval_steps_per_second": 3.272,
      "step": 860
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.3626731038093567,
      "learning_rate": 0.0001150392017106201,
      "loss": 0.9668,
      "step": 865
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.31055352091789246,
      "learning_rate": 0.00011397006414825373,
      "loss": 0.9272,
      "step": 870
    },
    {
      "epoch": 0.62,
      "grad_norm": 0.3563440144062042,
      "learning_rate": 0.00011290092658588738,
      "loss": 0.9574,
      "step": 875
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.44283735752105713,
      "learning_rate": 0.00011183178902352102,
      "loss": 0.9294,
      "step": 880
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.9413205981254578,
      "eval_runtime": 430.6763,
      "eval_samples_per_second": 26.073,
      "eval_steps_per_second": 3.26,
      "step": 880
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.41935306787490845,
      "learning_rate": 0.00011076265146115467,
      "loss": 0.9551,
      "step": 885
    },
    {
      "epoch": 0.63,
      "grad_norm": 0.3747510612010956,
      "learning_rate": 0.00010969351389878829,
      "loss": 0.9838,
      "step": 890
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.343944251537323,
      "learning_rate": 0.00010862437633642194,
      "loss": 0.9927,
      "step": 895
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.4465276598930359,
      "learning_rate": 0.00010755523877405558,
      "loss": 1.0054,
      "step": 900
    },
    {
      "epoch": 0.64,
      "eval_loss": 0.9400754570960999,
      "eval_runtime": 428.8696,
      "eval_samples_per_second": 26.183,
      "eval_steps_per_second": 3.274,
      "step": 900
    },
    {
      "epoch": 0.64,
      "grad_norm": 0.3230556845664978,
      "learning_rate": 0.00010648610121168923,
      "loss": 0.9675,
      "step": 905
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.384182333946228,
      "learning_rate": 0.00010541696364932286,
      "loss": 0.9806,
      "step": 910
    },
    {
      "epoch": 0.65,
      "grad_norm": 0.3371874690055847,
      "learning_rate": 0.00010434782608695651,
      "loss": 0.9508,
      "step": 915
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.3343549370765686,
      "learning_rate": 0.00010327868852459015,
      "loss": 0.964,
      "step": 920
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.9393041729927063,
      "eval_runtime": 429.8253,
      "eval_samples_per_second": 26.125,
      "eval_steps_per_second": 3.266,
      "step": 920
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.35873135924339294,
      "learning_rate": 0.0001022095509622238,
      "loss": 0.9702,
      "step": 925
    },
    {
      "epoch": 0.66,
      "grad_norm": 0.34200361371040344,
      "learning_rate": 0.00010114041339985743,
      "loss": 0.9559,
      "step": 930
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.37706607580184937,
      "learning_rate": 0.00010007127583749108,
      "loss": 0.9392,
      "step": 935
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.37498244643211365,
      "learning_rate": 9.900213827512472e-05,
      "loss": 0.8863,
      "step": 940
    },
    {
      "epoch": 0.67,
      "eval_loss": 0.9388808608055115,
      "eval_runtime": 428.8283,
      "eval_samples_per_second": 26.185,
      "eval_steps_per_second": 3.274,
      "step": 940
    },
    {
      "epoch": 0.67,
      "grad_norm": 0.29014042019844055,
      "learning_rate": 9.793300071275837e-05,
      "loss": 0.9793,
      "step": 945
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.5693466663360596,
      "learning_rate": 9.686386315039202e-05,
      "loss": 1.0071,
      "step": 950
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.3804801106452942,
      "learning_rate": 9.579472558802566e-05,
      "loss": 1.0006,
      "step": 955
    },
    {
      "epoch": 0.68,
      "grad_norm": 0.3488084673881531,
      "learning_rate": 9.47255880256593e-05,
      "loss": 0.9766,
      "step": 960
    },
    {
      "epoch": 0.68,
      "eval_loss": 0.9387741684913635,
      "eval_runtime": 430.4724,
      "eval_samples_per_second": 26.085,
      "eval_steps_per_second": 3.262,
      "step": 960
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.3514138460159302,
      "learning_rate": 9.365645046329294e-05,
      "loss": 0.9362,
      "step": 965
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.2945019602775574,
      "learning_rate": 9.258731290092659e-05,
      "loss": 0.9354,
      "step": 970
    },
    {
      "epoch": 0.69,
      "grad_norm": 0.41364601254463196,
      "learning_rate": 9.151817533856021e-05,
      "loss": 0.9893,
      "step": 975
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.3566943109035492,
      "learning_rate": 9.044903777619385e-05,
      "loss": 0.9803,
      "step": 980
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.9385978579521179,
      "eval_runtime": 429.3426,
      "eval_samples_per_second": 26.154,
      "eval_steps_per_second": 3.27,
      "step": 980
    },
    {
      "epoch": 0.7,
      "grad_norm": 0.3117084503173828,
      "learning_rate": 8.93799002138275e-05,
      "loss": 0.9584,
      "step": 985
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.3447965085506439,
      "learning_rate": 8.831076265146115e-05,
      "loss": 0.9283,
      "step": 990
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.43709927797317505,
      "learning_rate": 8.724162508909478e-05,
      "loss": 0.9653,
      "step": 995
    },
    {
      "epoch": 0.71,
      "grad_norm": 0.30693626403808594,
      "learning_rate": 8.617248752672843e-05,
      "loss": 0.9367,
      "step": 1000
    },
    {
      "epoch": 0.71,
      "eval_loss": 0.938434898853302,
      "eval_runtime": 429.8587,
      "eval_samples_per_second": 26.123,
      "eval_steps_per_second": 3.266,
      "step": 1000
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.3513086140155792,
      "learning_rate": 8.510334996436207e-05,
      "loss": 0.9486,
      "step": 1005
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.30035388469696045,
      "learning_rate": 8.403421240199572e-05,
      "loss": 0.9253,
      "step": 1010
    },
    {
      "epoch": 0.72,
      "grad_norm": 0.34532248973846436,
      "learning_rate": 8.296507483962936e-05,
      "loss": 0.9817,
      "step": 1015
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.2752978205680847,
      "learning_rate": 8.1895937277263e-05,
      "loss": 0.9092,
      "step": 1020
    },
    {
      "epoch": 0.73,
      "eval_loss": 0.9372485876083374,
      "eval_runtime": 429.979,
      "eval_samples_per_second": 26.115,
      "eval_steps_per_second": 3.265,
      "step": 1020
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.37310290336608887,
      "learning_rate": 8.082679971489664e-05,
      "loss": 0.9679,
      "step": 1025
    },
    {
      "epoch": 0.73,
      "grad_norm": 0.26144835352897644,
      "learning_rate": 7.975766215253029e-05,
      "loss": 0.9374,
      "step": 1030
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.442670613527298,
      "learning_rate": 7.868852459016393e-05,
      "loss": 0.9386,
      "step": 1035
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.3123500347137451,
      "learning_rate": 7.761938702779758e-05,
      "loss": 0.9527,
      "step": 1040
    },
    {
      "epoch": 0.74,
      "eval_loss": 0.9374279975891113,
      "eval_runtime": 428.7356,
      "eval_samples_per_second": 26.191,
      "eval_steps_per_second": 3.275,
      "step": 1040
    },
    {
      "epoch": 0.74,
      "grad_norm": 0.34689196944236755,
      "learning_rate": 7.655024946543121e-05,
      "loss": 0.918,
      "step": 1045
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.25890424847602844,
      "learning_rate": 7.548111190306486e-05,
      "loss": 0.9131,
      "step": 1050
    },
    {
      "epoch": 0.75,
      "grad_norm": 0.4108467102050781,
      "learning_rate": 7.44119743406985e-05,
      "loss": 0.9481,
      "step": 1055
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.3142194151878357,
      "learning_rate": 7.334283677833213e-05,
      "loss": 0.9503,
      "step": 1060
    },
    {
      "epoch": 0.76,
      "eval_loss": 0.9363455772399902,
      "eval_runtime": 429.8697,
      "eval_samples_per_second": 26.122,
      "eval_steps_per_second": 3.266,
      "step": 1060
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.3752140402793884,
      "learning_rate": 7.227369921596578e-05,
      "loss": 0.9614,
      "step": 1065
    },
    {
      "epoch": 0.76,
      "grad_norm": 0.3649095892906189,
      "learning_rate": 7.120456165359942e-05,
      "loss": 0.9428,
      "step": 1070
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.28237637877464294,
      "learning_rate": 7.013542409123307e-05,
      "loss": 0.9698,
      "step": 1075
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.32881519198417664,
      "learning_rate": 6.90662865288667e-05,
      "loss": 0.9494,
      "step": 1080
    },
    {
      "epoch": 0.77,
      "eval_loss": 0.9357610940933228,
      "eval_runtime": 430.0153,
      "eval_samples_per_second": 26.113,
      "eval_steps_per_second": 3.265,
      "step": 1080
    },
    {
      "epoch": 0.77,
      "grad_norm": 0.3415566086769104,
      "learning_rate": 6.799714896650034e-05,
      "loss": 0.9501,
      "step": 1085
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.3997558057308197,
      "learning_rate": 6.692801140413399e-05,
      "loss": 0.9423,
      "step": 1090
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.3917606472969055,
      "learning_rate": 6.585887384176763e-05,
      "loss": 0.9696,
      "step": 1095
    },
    {
      "epoch": 0.78,
      "grad_norm": 0.6566518545150757,
      "learning_rate": 6.478973627940128e-05,
      "loss": 0.961,
      "step": 1100
    },
    {
      "epoch": 0.78,
      "eval_loss": 0.936005711555481,
      "eval_runtime": 430.641,
      "eval_samples_per_second": 26.075,
      "eval_steps_per_second": 3.26,
      "step": 1100
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.2704831063747406,
      "learning_rate": 6.372059871703493e-05,
      "loss": 0.9316,
      "step": 1105
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.3626934885978699,
      "learning_rate": 6.265146115466856e-05,
      "loss": 0.942,
      "step": 1110
    },
    {
      "epoch": 0.79,
      "grad_norm": 0.3050808310508728,
      "learning_rate": 6.158232359230221e-05,
      "loss": 0.9681,
      "step": 1115
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.33575522899627686,
      "learning_rate": 6.051318602993584e-05,
      "loss": 0.9675,
      "step": 1120
    },
    {
      "epoch": 0.8,
      "eval_loss": 0.9356703758239746,
      "eval_runtime": 430.1349,
      "eval_samples_per_second": 26.106,
      "eval_steps_per_second": 3.264,
      "step": 1120
    },
    {
      "epoch": 0.8,
      "grad_norm": 0.47601962089538574,
      "learning_rate": 5.9444048467569485e-05,
      "loss": 0.9837,
      "step": 1125
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.3533668518066406,
      "learning_rate": 5.837491090520313e-05,
      "loss": 0.9234,
      "step": 1130
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.32863783836364746,
      "learning_rate": 5.730577334283677e-05,
      "loss": 0.9773,
      "step": 1135
    },
    {
      "epoch": 0.81,
      "grad_norm": 0.3788212835788727,
      "learning_rate": 5.6236635780470413e-05,
      "loss": 0.9704,
      "step": 1140
    },
    {
      "epoch": 0.81,
      "eval_loss": 0.9352059364318848,
      "eval_runtime": 430.5021,
      "eval_samples_per_second": 26.083,
      "eval_steps_per_second": 3.261,
      "step": 1140
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.5751201510429382,
      "learning_rate": 5.5167498218104056e-05,
      "loss": 0.9367,
      "step": 1145
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.4319859445095062,
      "learning_rate": 5.40983606557377e-05,
      "loss": 0.9558,
      "step": 1150
    },
    {
      "epoch": 0.82,
      "grad_norm": 0.35112935304641724,
      "learning_rate": 5.302922309337134e-05,
      "loss": 0.9123,
      "step": 1155
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.31719475984573364,
      "learning_rate": 5.196008553100499e-05,
      "loss": 0.9511,
      "step": 1160
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.9346606135368347,
      "eval_runtime": 429.4307,
      "eval_samples_per_second": 26.149,
      "eval_steps_per_second": 3.269,
      "step": 1160
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.3528451919555664,
      "learning_rate": 5.089094796863862e-05,
      "loss": 0.9789,
      "step": 1165
    },
    {
      "epoch": 0.83,
      "grad_norm": 0.29251405596733093,
      "learning_rate": 4.9821810406272264e-05,
      "loss": 0.9271,
      "step": 1170
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.3118594288825989,
      "learning_rate": 4.875267284390591e-05,
      "loss": 0.9356,
      "step": 1175
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.2919027805328369,
      "learning_rate": 4.7683535281539556e-05,
      "loss": 0.9042,
      "step": 1180
    },
    {
      "epoch": 0.84,
      "eval_loss": 0.9343792200088501,
      "eval_runtime": 429.6874,
      "eval_samples_per_second": 26.133,
      "eval_steps_per_second": 3.267,
      "step": 1180
    },
    {
      "epoch": 0.84,
      "grad_norm": 0.3041454553604126,
      "learning_rate": 4.66143977191732e-05,
      "loss": 0.9366,
      "step": 1185
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.42262208461761475,
      "learning_rate": 4.554526015680684e-05,
      "loss": 0.954,
      "step": 1190
    },
    {
      "epoch": 0.85,
      "grad_norm": 0.4994548559188843,
      "learning_rate": 4.4476122594440485e-05,
      "loss": 0.9125,
      "step": 1195
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.506973385810852,
      "learning_rate": 4.340698503207413e-05,
      "loss": 0.9359,
      "step": 1200
    },
    {
      "epoch": 0.86,
      "eval_loss": 0.9343673586845398,
      "eval_runtime": 429.7828,
      "eval_samples_per_second": 26.127,
      "eval_steps_per_second": 3.267,
      "step": 1200
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.3126988112926483,
      "learning_rate": 4.2337847469707764e-05,
      "loss": 0.9643,
      "step": 1205
    },
    {
      "epoch": 0.86,
      "grad_norm": 0.33700090646743774,
      "learning_rate": 4.1268709907341407e-05,
      "loss": 0.9809,
      "step": 1210
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.27363497018814087,
      "learning_rate": 4.019957234497505e-05,
      "loss": 0.9519,
      "step": 1215
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.2861486077308655,
      "learning_rate": 3.913043478260869e-05,
      "loss": 0.9131,
      "step": 1220
    },
    {
      "epoch": 0.87,
      "eval_loss": 0.9337981343269348,
      "eval_runtime": 429.8829,
      "eval_samples_per_second": 26.121,
      "eval_steps_per_second": 3.266,
      "step": 1220
    },
    {
      "epoch": 0.87,
      "grad_norm": 0.6844860911369324,
      "learning_rate": 3.8061297220242335e-05,
      "loss": 0.996,
      "step": 1225
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.4049902558326721,
      "learning_rate": 3.699215965787598e-05,
      "loss": 0.9593,
      "step": 1230
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.3516334295272827,
      "learning_rate": 3.592302209550962e-05,
      "loss": 0.9,
      "step": 1235
    },
    {
      "epoch": 0.88,
      "grad_norm": 0.3603789508342743,
      "learning_rate": 3.485388453314326e-05,
      "loss": 0.9674,
      "step": 1240
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.9333547353744507,
      "eval_runtime": 428.501,
      "eval_samples_per_second": 26.205,
      "eval_steps_per_second": 3.277,
      "step": 1240
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.3410665690898895,
      "learning_rate": 3.3784746970776906e-05,
      "loss": 0.9335,
      "step": 1245
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.34668412804603577,
      "learning_rate": 3.271560940841055e-05,
      "loss": 0.9618,
      "step": 1250
    },
    {
      "epoch": 0.89,
      "grad_norm": 0.361831396818161,
      "learning_rate": 3.164647184604419e-05,
      "loss": 0.9524,
      "step": 1255
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.4450233280658722,
      "learning_rate": 3.057733428367783e-05,
      "loss": 0.9855,
      "step": 1260
    },
    {
      "epoch": 0.9,
      "eval_loss": 0.9331775307655334,
      "eval_runtime": 430.3503,
      "eval_samples_per_second": 26.093,
      "eval_steps_per_second": 3.262,
      "step": 1260
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.38541775941848755,
      "learning_rate": 2.950819672131147e-05,
      "loss": 0.9691,
      "step": 1265
    },
    {
      "epoch": 0.9,
      "grad_norm": 0.43870338797569275,
      "learning_rate": 2.8439059158945114e-05,
      "loss": 0.9514,
      "step": 1270
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.3042528033256531,
      "learning_rate": 2.736992159657876e-05,
      "loss": 0.9099,
      "step": 1275
    },
    {
      "epoch": 0.91,
      "grad_norm": 0.28177541494369507,
      "learning_rate": 2.63007840342124e-05,
      "loss": 0.9201,
      "step": 1280
    },
    {
      "epoch": 0.91,
      "eval_loss": 0.9329990744590759,
      "eval_runtime": 430.5216,
      "eval_samples_per_second": 26.082,
      "eval_steps_per_second": 3.261,
      "step": 1280
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.31207114458084106,
      "learning_rate": 2.5231646471846042e-05,
      "loss": 0.9388,
      "step": 1285
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.39428967237472534,
      "learning_rate": 2.4162508909479685e-05,
      "loss": 0.9827,
      "step": 1290
    },
    {
      "epoch": 0.92,
      "grad_norm": 0.35494768619537354,
      "learning_rate": 2.3093371347113328e-05,
      "loss": 0.9094,
      "step": 1295
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.5314223766326904,
      "learning_rate": 2.2024233784746968e-05,
      "loss": 0.9453,
      "step": 1300
    },
    {
      "epoch": 0.93,
      "eval_loss": 0.9329606294631958,
      "eval_runtime": 429.558,
      "eval_samples_per_second": 26.141,
      "eval_steps_per_second": 3.268,
      "step": 1300
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.3589262366294861,
      "learning_rate": 2.095509622238061e-05,
      "loss": 0.9202,
      "step": 1305
    },
    {
      "epoch": 0.93,
      "grad_norm": 0.3077951967716217,
      "learning_rate": 1.9885958660014253e-05,
      "loss": 0.9741,
      "step": 1310
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.3348515033721924,
      "learning_rate": 1.8816821097647896e-05,
      "loss": 0.933,
      "step": 1315
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.34439706802368164,
      "learning_rate": 1.774768353528154e-05,
      "loss": 0.9476,
      "step": 1320
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.932703971862793,
      "eval_runtime": 428.8979,
      "eval_samples_per_second": 26.181,
      "eval_steps_per_second": 3.274,
      "step": 1320
    },
    {
      "epoch": 0.94,
      "grad_norm": 0.28818029165267944,
      "learning_rate": 1.6678545972915182e-05,
      "loss": 0.9288,
      "step": 1325
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.2997274696826935,
      "learning_rate": 1.560940841054882e-05,
      "loss": 0.9418,
      "step": 1330
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.3139705955982208,
      "learning_rate": 1.4540270848182466e-05,
      "loss": 0.9732,
      "step": 1335
    },
    {
      "epoch": 0.95,
      "grad_norm": 0.32881295680999756,
      "learning_rate": 1.3471133285816107e-05,
      "loss": 0.9166,
      "step": 1340
    },
    {
      "epoch": 0.95,
      "eval_loss": 0.9325681328773499,
      "eval_runtime": 429.5609,
      "eval_samples_per_second": 26.141,
      "eval_steps_per_second": 3.268,
      "step": 1340
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.6226952075958252,
      "learning_rate": 1.240199572344975e-05,
      "loss": 1.0029,
      "step": 1345
    },
    {
      "epoch": 0.96,
      "grad_norm": 0.3391931354999542,
      "learning_rate": 1.1332858161083391e-05,
      "loss": 0.9522,
      "step": 1350
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.37440812587738037,
      "learning_rate": 1.0263720598717034e-05,
      "loss": 0.9362,
      "step": 1355
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.3198889493942261,
      "learning_rate": 9.194583036350677e-06,
      "loss": 0.9523,
      "step": 1360
    },
    {
      "epoch": 0.97,
      "eval_loss": 0.9325061440467834,
      "eval_runtime": 429.5538,
      "eval_samples_per_second": 26.141,
      "eval_steps_per_second": 3.269,
      "step": 1360
    },
    {
      "epoch": 0.97,
      "grad_norm": 0.4025633931159973,
      "learning_rate": 8.12544547398432e-06,
      "loss": 0.9422,
      "step": 1365
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.38235458731651306,
      "learning_rate": 7.0563079116179615e-06,
      "loss": 1.0052,
      "step": 1370
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.3941135108470917,
      "learning_rate": 5.9871703492516035e-06,
      "loss": 0.9714,
      "step": 1375
    },
    {
      "epoch": 0.98,
      "grad_norm": 0.31835320591926575,
      "learning_rate": 4.9180327868852455e-06,
      "loss": 0.9615,
      "step": 1380
    },
    {
      "epoch": 0.98,
      "eval_loss": 0.9324172735214233,
      "eval_runtime": 431.2711,
      "eval_samples_per_second": 26.037,
      "eval_steps_per_second": 3.255,
      "step": 1380
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.3852035105228424,
      "learning_rate": 3.848895224518888e-06,
      "loss": 0.9238,
      "step": 1385
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.33982229232788086,
      "learning_rate": 2.7797576621525303e-06,
      "loss": 0.9401,
      "step": 1390
    },
    {
      "epoch": 0.99,
      "grad_norm": 0.31724292039871216,
      "learning_rate": 1.7106200997861725e-06,
      "loss": 0.9274,
      "step": 1395
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.35556310415267944,
      "learning_rate": 6.414825374198146e-07,
      "loss": 0.9633,
      "step": 1400
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.9323536157608032,
      "eval_runtime": 430.5102,
      "eval_samples_per_second": 26.083,
      "eval_steps_per_second": 3.261,
      "step": 1400
    }
  ],
  "logging_steps": 5,
  "max_steps": 1403,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 20,
  "total_flos": 5.2878011087683584e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}