{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 29.454746136865342,
  "global_step": 20000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.59,
      "learning_rate": 0.0002382,
      "loss": 6.1211,
      "step": 400
    },
    {
      "epoch": 0.59,
      "eval_loss": 3.2328364849090576,
      "eval_runtime": 546.0856,
      "eval_samples_per_second": 13.958,
      "eval_steps_per_second": 1.745,
      "eval_wer": 0.9991302123305194,
      "step": 400
    },
    {
      "epoch": 1.18,
      "learning_rate": 0.0002955158530447911,
      "loss": 2.304,
      "step": 800
    },
    {
      "epoch": 1.18,
      "eval_loss": 1.0820916891098022,
      "eval_runtime": 551.5198,
      "eval_samples_per_second": 13.82,
      "eval_steps_per_second": 1.728,
      "eval_wer": 0.8043745203376823,
      "step": 800
    },
    {
      "epoch": 1.77,
      "learning_rate": 0.0002894765978862607,
      "loss": 0.9673,
      "step": 1200
    },
    {
      "epoch": 1.77,
      "eval_loss": 0.7271230816841125,
      "eval_runtime": 547.9451,
      "eval_samples_per_second": 13.91,
      "eval_steps_per_second": 1.739,
      "eval_wer": 0.6917881811204911,
      "step": 1200
    },
    {
      "epoch": 2.36,
      "learning_rate": 0.0002834373427277302,
      "loss": 0.7607,
      "step": 1600
    },
    {
      "epoch": 2.36,
      "eval_loss": 0.6257076263427734,
      "eval_runtime": 545.1866,
      "eval_samples_per_second": 13.981,
      "eval_steps_per_second": 1.748,
      "eval_wer": 0.6329751854694295,
      "step": 1600
    },
    {
      "epoch": 2.94,
      "learning_rate": 0.00027739808756919976,
      "loss": 0.689,
      "step": 2000
    },
    {
      "epoch": 2.94,
      "eval_loss": 0.5595377683639526,
      "eval_runtime": 545.7121,
      "eval_samples_per_second": 13.967,
      "eval_steps_per_second": 1.746,
      "eval_wer": 0.6035303146584804,
      "step": 2000
    },
    {
      "epoch": 3.53,
      "learning_rate": 0.00027135883241066934,
      "loss": 0.5775,
      "step": 2400
    },
    {
      "epoch": 3.53,
      "eval_loss": 0.582660973072052,
      "eval_runtime": 547.0999,
      "eval_samples_per_second": 13.932,
      "eval_steps_per_second": 1.742,
      "eval_wer": 0.6055257099002302,
      "step": 2400
    },
    {
      "epoch": 4.12,
      "learning_rate": 0.00026531957725213886,
      "loss": 0.5621,
      "step": 2800
    },
    {
      "epoch": 4.12,
      "eval_loss": 0.5549562573432922,
      "eval_runtime": 546.696,
      "eval_samples_per_second": 13.942,
      "eval_steps_per_second": 1.743,
      "eval_wer": 0.5691992837042722,
      "step": 2800
    },
    {
      "epoch": 4.71,
      "learning_rate": 0.00025928032209360843,
      "loss": 0.5014,
      "step": 3200
    },
    {
      "epoch": 4.71,
      "eval_loss": 0.5490128993988037,
      "eval_runtime": 542.4177,
      "eval_samples_per_second": 14.052,
      "eval_steps_per_second": 1.757,
      "eval_wer": 0.5637503197748784,
      "step": 3200
    },
    {
      "epoch": 5.3,
      "learning_rate": 0.000253241066935078,
      "loss": 0.4781,
      "step": 3600
    },
    {
      "epoch": 5.3,
      "eval_loss": 0.5757995247840881,
      "eval_runtime": 543.8318,
      "eval_samples_per_second": 14.015,
      "eval_steps_per_second": 1.752,
      "eval_wer": 0.5655154771041186,
      "step": 3600
    },
    {
      "epoch": 5.89,
      "learning_rate": 0.0002472018117765476,
      "loss": 0.4499,
      "step": 4000
    },
    {
      "epoch": 5.89,
      "eval_loss": 0.5555837154388428,
      "eval_runtime": 545.7873,
      "eval_samples_per_second": 13.965,
      "eval_steps_per_second": 1.746,
      "eval_wer": 0.5522640061396776,
      "step": 4000
    },
    {
      "epoch": 6.48,
      "learning_rate": 0.0002411625566180171,
      "loss": 0.4095,
      "step": 4400
    },
    {
      "epoch": 6.48,
      "eval_loss": 0.5786208510398865,
      "eval_runtime": 543.989,
      "eval_samples_per_second": 14.011,
      "eval_steps_per_second": 1.752,
      "eval_wer": 0.5524942440521873,
      "step": 4400
    },
    {
      "epoch": 7.07,
      "learning_rate": 0.00023512330145948666,
      "loss": 0.4003,
      "step": 4800
    },
    {
      "epoch": 7.07,
      "eval_loss": 0.5860427021980286,
      "eval_runtime": 549.6883,
      "eval_samples_per_second": 13.866,
      "eval_steps_per_second": 1.734,
      "eval_wer": 0.5389613711946789,
      "step": 4800
    },
    {
      "epoch": 7.66,
      "learning_rate": 0.00022908404630095618,
      "loss": 0.3653,
      "step": 5200
    },
    {
      "epoch": 7.66,
      "eval_loss": 0.5734272003173828,
      "eval_runtime": 542.4405,
      "eval_samples_per_second": 14.051,
      "eval_steps_per_second": 1.757,
      "eval_wer": 0.5064466615502686,
      "step": 5200
    },
    {
      "epoch": 8.25,
      "learning_rate": 0.00022304479114242576,
      "loss": 0.3454,
      "step": 5600
    },
    {
      "epoch": 8.25,
      "eval_loss": 0.5864331126213074,
      "eval_runtime": 543.9245,
      "eval_samples_per_second": 14.013,
      "eval_steps_per_second": 1.752,
      "eval_wer": 0.4945510360706063,
      "step": 5600
    },
    {
      "epoch": 8.84,
      "learning_rate": 0.0002170055359838953,
      "loss": 0.3223,
      "step": 6000
    },
    {
      "epoch": 8.84,
      "eval_loss": 0.5884710550308228,
      "eval_runtime": 548.5513,
      "eval_samples_per_second": 13.895,
      "eval_steps_per_second": 1.737,
      "eval_wer": 0.500332565873625,
      "step": 6000
    },
    {
      "epoch": 9.43,
      "learning_rate": 0.00021096628082536487,
      "loss": 0.2897,
      "step": 6400
    },
    {
      "epoch": 9.43,
      "eval_loss": 0.6017025113105774,
      "eval_runtime": 545.222,
      "eval_samples_per_second": 13.98,
      "eval_steps_per_second": 1.748,
      "eval_wer": 0.49053466359682785,
      "step": 6400
    },
    {
      "epoch": 10.01,
      "learning_rate": 0.0002049270256668344,
      "loss": 0.289,
      "step": 6800
    },
    {
      "epoch": 10.01,
      "eval_loss": 0.63252192735672,
      "eval_runtime": 547.5169,
      "eval_samples_per_second": 13.921,
      "eval_steps_per_second": 1.741,
      "eval_wer": 0.4930928626247122,
      "step": 6800
    },
    {
      "epoch": 10.6,
      "learning_rate": 0.00019888777050830396,
      "loss": 0.2488,
      "step": 7200
    },
    {
      "epoch": 10.6,
      "eval_loss": 0.6799584627151489,
      "eval_runtime": 541.9978,
      "eval_samples_per_second": 14.063,
      "eval_steps_per_second": 1.758,
      "eval_wer": 0.49833717063187516,
      "step": 7200
    },
    {
      "epoch": 11.19,
      "learning_rate": 0.0001928485153497735,
      "loss": 0.2361,
      "step": 7600
    },
    {
      "epoch": 11.19,
      "eval_loss": 0.6552415490150452,
      "eval_runtime": 544.0722,
      "eval_samples_per_second": 14.009,
      "eval_steps_per_second": 1.752,
      "eval_wer": 0.5002046559222307,
      "step": 7600
    },
    {
      "epoch": 11.78,
      "learning_rate": 0.00018682435832913938,
      "loss": 0.2275,
      "step": 8000
    },
    {
      "epoch": 11.78,
      "eval_loss": 0.6828446984291077,
      "eval_runtime": 543.1333,
      "eval_samples_per_second": 14.033,
      "eval_steps_per_second": 1.755,
      "eval_wer": 0.4898183678690202,
      "step": 8000
    },
    {
      "epoch": 12.37,
      "learning_rate": 0.00018078510317060895,
      "loss": 0.2109,
      "step": 8400
    },
    {
      "epoch": 12.37,
      "eval_loss": 0.6952915787696838,
      "eval_runtime": 542.929,
      "eval_samples_per_second": 14.039,
      "eval_steps_per_second": 1.755,
      "eval_wer": 0.4861601432591456,
      "step": 8400
    },
    {
      "epoch": 12.96,
      "learning_rate": 0.0001747458480120785,
      "loss": 0.2061,
      "step": 8800
    },
    {
      "epoch": 12.96,
      "eval_loss": 0.6886131167411804,
      "eval_runtime": 544.2666,
      "eval_samples_per_second": 14.004,
      "eval_steps_per_second": 1.751,
      "eval_wer": 0.4692504476848299,
      "step": 8800
    },
    {
      "epoch": 13.55,
      "learning_rate": 0.00016870659285354804,
      "loss": 0.1874,
      "step": 9200
    },
    {
      "epoch": 13.55,
      "eval_loss": 0.7013294696807861,
      "eval_runtime": 543.7236,
      "eval_samples_per_second": 14.018,
      "eval_steps_per_second": 1.753,
      "eval_wer": 0.47367613200306985,
      "step": 9200
    },
    {
      "epoch": 14.14,
      "learning_rate": 0.00016266733769501759,
      "loss": 0.1824,
      "step": 9600
    },
    {
      "epoch": 14.14,
      "eval_loss": 0.7199532985687256,
      "eval_runtime": 545.3665,
      "eval_samples_per_second": 13.976,
      "eval_steps_per_second": 1.747,
      "eval_wer": 0.4730621642363776,
      "step": 9600
    },
    {
      "epoch": 14.73,
      "learning_rate": 0.00015662808253648716,
      "loss": 0.1773,
      "step": 10000
    },
    {
      "epoch": 14.73,
      "eval_loss": 0.6803578734397888,
      "eval_runtime": 541.6325,
      "eval_samples_per_second": 14.072,
      "eval_steps_per_second": 1.759,
      "eval_wer": 0.47045280122793554,
      "step": 10000
    },
    {
      "epoch": 15.32,
      "learning_rate": 0.00015058882737795668,
      "loss": 0.1663,
      "step": 10400
    },
    {
      "epoch": 15.32,
      "eval_loss": 0.6929047703742981,
      "eval_runtime": 543.7575,
      "eval_samples_per_second": 14.017,
      "eval_steps_per_second": 1.753,
      "eval_wer": 0.4615246866206191,
      "step": 10400
    },
    {
      "epoch": 15.91,
      "learning_rate": 0.00014454957221942625,
      "loss": 0.1529,
      "step": 10800
    },
    {
      "epoch": 15.91,
      "eval_loss": 0.7400447130203247,
      "eval_runtime": 541.8998,
      "eval_samples_per_second": 14.065,
      "eval_steps_per_second": 1.759,
      "eval_wer": 0.4675364543361474,
      "step": 10800
    },
    {
      "epoch": 16.49,
      "learning_rate": 0.00013851031706089582,
      "loss": 0.1406,
      "step": 11200
    },
    {
      "epoch": 16.49,
      "eval_loss": 0.7907389998435974,
      "eval_runtime": 543.7393,
      "eval_samples_per_second": 14.018,
      "eval_steps_per_second": 1.753,
      "eval_wer": 0.46472243540547453,
      "step": 11200
    },
    {
      "epoch": 17.08,
      "learning_rate": 0.0001324861600402617,
      "loss": 0.1376,
      "step": 11600
    },
    {
      "epoch": 17.08,
      "eval_loss": 0.8006933927536011,
      "eval_runtime": 543.2472,
      "eval_samples_per_second": 14.03,
      "eval_steps_per_second": 1.754,
      "eval_wer": 0.46689690457917626,
      "step": 11600
    },
    {
      "epoch": 17.67,
      "learning_rate": 0.00012644690488173123,
      "loss": 0.1273,
      "step": 12000
    },
    {
      "epoch": 17.67,
      "eval_loss": 0.7544116973876953,
      "eval_runtime": 544.2672,
      "eval_samples_per_second": 14.004,
      "eval_steps_per_second": 1.751,
      "eval_wer": 0.46403172166794576,
      "step": 12000
    },
    {
      "epoch": 18.26,
      "learning_rate": 0.00012040764972320079,
      "loss": 0.1252,
      "step": 12400
    },
    {
      "epoch": 18.26,
      "eval_loss": 0.7654944658279419,
      "eval_runtime": 548.3671,
      "eval_samples_per_second": 13.899,
      "eval_steps_per_second": 1.738,
      "eval_wer": 0.45349194167306217,
      "step": 12400
    },
    {
      "epoch": 18.85,
      "learning_rate": 0.00011436839456467034,
      "loss": 0.1226,
      "step": 12800
    },
    {
      "epoch": 18.85,
      "eval_loss": 0.7361114621162415,
      "eval_runtime": 546.1341,
      "eval_samples_per_second": 13.956,
      "eval_steps_per_second": 1.745,
      "eval_wer": 0.4549245331286774,
      "step": 12800
    },
    {
      "epoch": 19.44,
      "learning_rate": 0.0001083291394061399,
      "loss": 0.1165,
      "step": 13200
    },
    {
      "epoch": 19.44,
      "eval_loss": 0.8643974661827087,
      "eval_runtime": 543.4889,
      "eval_samples_per_second": 14.024,
      "eval_steps_per_second": 1.753,
      "eval_wer": 0.45622921463289845,
      "step": 13200
    },
    {
      "epoch": 20.03,
      "learning_rate": 0.00010230498238550579,
      "loss": 0.1119,
      "step": 13600
    },
    {
      "epoch": 20.03,
      "eval_loss": 0.8597950339317322,
      "eval_runtime": 545.2818,
      "eval_samples_per_second": 13.978,
      "eval_steps_per_second": 1.748,
      "eval_wer": 0.4632642619595805,
      "step": 13600
    },
    {
      "epoch": 20.62,
      "learning_rate": 9.626572722697534e-05,
      "loss": 0.1077,
      "step": 14000
    },
    {
      "epoch": 20.62,
      "eval_loss": 0.8093447089195251,
      "eval_runtime": 541.9895,
      "eval_samples_per_second": 14.063,
      "eval_steps_per_second": 1.758,
      "eval_wer": 0.453031465848043,
      "step": 14000
    },
    {
      "epoch": 21.21,
      "learning_rate": 9.02264720684449e-05,
      "loss": 0.102,
      "step": 14400
    },
    {
      "epoch": 21.21,
      "eval_loss": 0.8589721918106079,
      "eval_runtime": 548.1684,
      "eval_samples_per_second": 13.904,
      "eval_steps_per_second": 1.739,
      "eval_wer": 0.4514965464313124,
      "step": 14400
    },
    {
      "epoch": 21.8,
      "learning_rate": 8.418721690991444e-05,
      "loss": 0.0939,
      "step": 14800
    },
    {
      "epoch": 21.8,
      "eval_loss": 0.8175553679466248,
      "eval_runtime": 541.1705,
      "eval_samples_per_second": 14.084,
      "eval_steps_per_second": 1.761,
      "eval_wer": 0.4527500639549757,
      "step": 14800
    },
    {
      "epoch": 22.39,
      "learning_rate": 7.814796175138399e-05,
      "loss": 0.0898,
      "step": 15200
    },
    {
      "epoch": 22.39,
      "eval_loss": 0.8843649625778198,
      "eval_runtime": 546.2264,
      "eval_samples_per_second": 13.954,
      "eval_steps_per_second": 1.745,
      "eval_wer": 0.44740342798669736,
      "step": 15200
    },
    {
      "epoch": 22.97,
      "learning_rate": 7.210870659285354e-05,
      "loss": 0.0903,
      "step": 15600
    },
    {
      "epoch": 22.97,
      "eval_loss": 0.8875829577445984,
      "eval_runtime": 540.326,
      "eval_samples_per_second": 14.106,
      "eval_steps_per_second": 1.764,
      "eval_wer": 0.45231517012023537,
      "step": 15600
    },
    {
      "epoch": 23.56,
      "learning_rate": 6.606945143432309e-05,
      "loss": 0.0848,
      "step": 16000
    },
    {
      "epoch": 23.56,
      "eval_loss": 0.9255176782608032,
      "eval_runtime": 544.8055,
      "eval_samples_per_second": 13.99,
      "eval_steps_per_second": 1.749,
      "eval_wer": 0.44814530570478384,
      "step": 16000
    },
    {
      "epoch": 24.15,
      "learning_rate": 6.003019627579265e-05,
      "loss": 0.0822,
      "step": 16400
    },
    {
      "epoch": 24.15,
      "eval_loss": 0.9284627437591553,
      "eval_runtime": 544.7798,
      "eval_samples_per_second": 13.991,
      "eval_steps_per_second": 1.749,
      "eval_wer": 0.4471220260936301,
      "step": 16400
    },
    {
      "epoch": 24.74,
      "learning_rate": 5.39909411172622e-05,
      "loss": 0.0767,
      "step": 16800
    },
    {
      "epoch": 24.74,
      "eval_loss": 0.9410629868507385,
      "eval_runtime": 544.0954,
      "eval_samples_per_second": 14.009,
      "eval_steps_per_second": 1.752,
      "eval_wer": 0.4394218470196981,
      "step": 16800
    },
    {
      "epoch": 25.33,
      "learning_rate": 4.795168595873175e-05,
      "loss": 0.0735,
      "step": 17200
    },
    {
      "epoch": 25.33,
      "eval_loss": 0.9868486523628235,
      "eval_runtime": 546.6128,
      "eval_samples_per_second": 13.944,
      "eval_steps_per_second": 1.743,
      "eval_wer": 0.44195446405730365,
      "step": 17200
    },
    {
      "epoch": 25.92,
      "learning_rate": 4.1927528938097633e-05,
      "loss": 0.0728,
      "step": 17600
    },
    {
      "epoch": 25.92,
      "eval_loss": 0.938178539276123,
      "eval_runtime": 542.4257,
      "eval_samples_per_second": 14.052,
      "eval_steps_per_second": 1.757,
      "eval_wer": 0.4406753645433615,
      "step": 17600
    },
    {
      "epoch": 26.51,
      "learning_rate": 3.590337191746351e-05,
      "loss": 0.0673,
      "step": 18000
    },
    {
      "epoch": 26.51,
      "eval_loss": 0.9842382669448853,
      "eval_runtime": 544.7728,
      "eval_samples_per_second": 13.991,
      "eval_steps_per_second": 1.749,
      "eval_wer": 0.4360706062931696,
      "step": 18000
    },
    {
      "epoch": 27.1,
      "learning_rate": 2.9864116758933062e-05,
      "loss": 0.0683,
      "step": 18400
    },
    {
      "epoch": 27.1,
      "eval_loss": 0.9531042575836182,
      "eval_runtime": 545.0114,
      "eval_samples_per_second": 13.985,
      "eval_steps_per_second": 1.749,
      "eval_wer": 0.4396520849322077,
      "step": 18400
    },
    {
      "epoch": 27.69,
      "learning_rate": 2.3824861600402614e-05,
      "loss": 0.0631,
      "step": 18800
    },
    {
      "epoch": 27.69,
      "eval_loss": 0.9613842964172363,
      "eval_runtime": 542.3788,
      "eval_samples_per_second": 14.053,
      "eval_steps_per_second": 1.757,
      "eval_wer": 0.4402148887183423,
      "step": 18800
    },
    {
      "epoch": 28.28,
      "learning_rate": 1.7785606441872167e-05,
      "loss": 0.0625,
      "step": 19200
    },
    {
      "epoch": 28.28,
      "eval_loss": 0.9772672057151794,
      "eval_runtime": 545.7524,
      "eval_samples_per_second": 13.966,
      "eval_steps_per_second": 1.746,
      "eval_wer": 0.4349194167306216,
      "step": 19200
    },
    {
      "epoch": 28.87,
      "learning_rate": 1.174635128334172e-05,
      "loss": 0.0599,
      "step": 19600
    },
    {
      "epoch": 28.87,
      "eval_loss": 0.9910905361175537,
      "eval_runtime": 543.4098,
      "eval_samples_per_second": 14.026,
      "eval_steps_per_second": 1.754,
      "eval_wer": 0.4378613456126887,
      "step": 19600
    },
    {
      "epoch": 29.45,
      "learning_rate": 5.707096124811273e-06,
      "loss": 0.0576,
      "step": 20000
    },
    {
      "epoch": 29.45,
      "eval_loss": 0.9930649995803833,
      "eval_runtime": 544.5195,
      "eval_samples_per_second": 13.998,
      "eval_steps_per_second": 1.75,
      "eval_wer": 0.435354310565362,
      "step": 20000
    }
  ],
  "max_steps": 20370,
  "num_train_epochs": 30,
  "total_flos": 8.26011084747878e+19,
  "trial_name": null,
  "trial_params": null
}