{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 22995,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 4.9648474306008555e-05,
      "loss": 4.8007,
      "step": 100
    },
    {
      "epoch": 0.04,
      "learning_rate": 4.9286076683336956e-05,
      "loss": 4.0366,
      "step": 200
    },
    {
      "epoch": 0.07,
      "learning_rate": 4.892367906066536e-05,
      "loss": 3.8316,
      "step": 300
    },
    {
      "epoch": 0.09,
      "learning_rate": 4.8561281437993764e-05,
      "loss": 3.689,
      "step": 400
    },
    {
      "epoch": 0.11,
      "learning_rate": 4.819888381532217e-05,
      "loss": 3.5765,
      "step": 500
    },
    {
      "epoch": 0.11,
      "eval_cer": 10.3648,
      "eval_gen_len": 16.358,
      "eval_loss": 3.433837413787842,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 38.4477,
      "eval_samples_per_second": 52.019,
      "eval_steps_per_second": 0.208,
      "step": 500
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.783648619265057e-05,
      "loss": 3.484,
      "step": 600
    },
    {
      "epoch": 0.15,
      "learning_rate": 4.747408856997898e-05,
      "loss": 3.3864,
      "step": 700
    },
    {
      "epoch": 0.17,
      "learning_rate": 4.711169094730739e-05,
      "loss": 3.3142,
      "step": 800
    },
    {
      "epoch": 0.2,
      "learning_rate": 4.674929332463579e-05,
      "loss": 3.2487,
      "step": 900
    },
    {
      "epoch": 0.22,
      "learning_rate": 4.63868957019642e-05,
      "loss": 3.1749,
      "step": 1000
    },
    {
      "epoch": 0.22,
      "eval_cer": 10.955,
      "eval_gen_len": 15.828,
      "eval_loss": 3.0109550952911377,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 38.0142,
      "eval_samples_per_second": 52.612,
      "eval_steps_per_second": 0.21,
      "step": 1000
    },
    {
      "epoch": 0.24,
      "learning_rate": 4.60244980792926e-05,
      "loss": 3.1153,
      "step": 1100
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.5662100456621006e-05,
      "loss": 3.0549,
      "step": 1200
    },
    {
      "epoch": 0.28,
      "learning_rate": 4.529970283394941e-05,
      "loss": 2.9852,
      "step": 1300
    },
    {
      "epoch": 0.3,
      "learning_rate": 4.4937305211277815e-05,
      "loss": 2.9315,
      "step": 1400
    },
    {
      "epoch": 0.33,
      "learning_rate": 4.4574907588606215e-05,
      "loss": 2.8696,
      "step": 1500
    },
    {
      "epoch": 0.33,
      "eval_cer": 11.2579,
      "eval_gen_len": 15.358,
      "eval_loss": 2.6943440437316895,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 36.6296,
      "eval_samples_per_second": 54.601,
      "eval_steps_per_second": 0.218,
      "step": 1500
    },
    {
      "epoch": 0.35,
      "learning_rate": 4.421250996593462e-05,
      "loss": 2.8218,
      "step": 1600
    },
    {
      "epoch": 0.37,
      "learning_rate": 4.385011234326303e-05,
      "loss": 2.7759,
      "step": 1700
    },
    {
      "epoch": 0.39,
      "learning_rate": 4.348771472059143e-05,
      "loss": 2.7254,
      "step": 1800
    },
    {
      "epoch": 0.41,
      "learning_rate": 4.312531709791984e-05,
      "loss": 2.6728,
      "step": 1900
    },
    {
      "epoch": 0.43,
      "learning_rate": 4.276291947524824e-05,
      "loss": 2.6279,
      "step": 2000
    },
    {
      "epoch": 0.43,
      "eval_cer": 11.6734,
      "eval_gen_len": 14.8565,
      "eval_loss": 2.4299662113189697,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 35.0111,
      "eval_samples_per_second": 57.125,
      "eval_steps_per_second": 0.228,
      "step": 2000
    },
    {
      "epoch": 0.46,
      "learning_rate": 4.240052185257665e-05,
      "loss": 2.5848,
      "step": 2100
    },
    {
      "epoch": 0.48,
      "learning_rate": 4.203812422990505e-05,
      "loss": 2.5407,
      "step": 2200
    },
    {
      "epoch": 0.5,
      "learning_rate": 4.167572660723346e-05,
      "loss": 2.4987,
      "step": 2300
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.131332898456186e-05,
      "loss": 2.4587,
      "step": 2400
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.0950931361890266e-05,
      "loss": 2.4166,
      "step": 2500
    },
    {
      "epoch": 0.54,
      "eval_cer": 11.028,
      "eval_gen_len": 14.523,
      "eval_loss": 2.224823474884033,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 32.1586,
      "eval_samples_per_second": 62.192,
      "eval_steps_per_second": 0.249,
      "step": 2500
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.058853373921867e-05,
      "loss": 2.3746,
      "step": 2600
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.0226136116547074e-05,
      "loss": 2.3435,
      "step": 2700
    },
    {
      "epoch": 0.61,
      "learning_rate": 3.986373849387548e-05,
      "loss": 2.3134,
      "step": 2800
    },
    {
      "epoch": 0.63,
      "learning_rate": 3.950134087120388e-05,
      "loss": 2.2865,
      "step": 2900
    },
    {
      "epoch": 0.65,
      "learning_rate": 3.913894324853229e-05,
      "loss": 2.2444,
      "step": 3000
    },
    {
      "epoch": 0.65,
      "eval_cer": 10.4379,
      "eval_gen_len": 14.4185,
      "eval_loss": 2.0489799976348877,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 32.7499,
      "eval_samples_per_second": 61.069,
      "eval_steps_per_second": 0.244,
      "step": 3000
    },
    {
      "epoch": 0.67,
      "learning_rate": 3.877654562586069e-05,
      "loss": 2.2154,
      "step": 3100
    },
    {
      "epoch": 0.7,
      "learning_rate": 3.84141480031891e-05,
      "loss": 2.1757,
      "step": 3200
    },
    {
      "epoch": 0.72,
      "learning_rate": 3.80517503805175e-05,
      "loss": 2.1584,
      "step": 3300
    },
    {
      "epoch": 0.74,
      "learning_rate": 3.768935275784591e-05,
      "loss": 2.1166,
      "step": 3400
    },
    {
      "epoch": 0.76,
      "learning_rate": 3.7326955135174316e-05,
      "loss": 2.0902,
      "step": 3500
    },
    {
      "epoch": 0.76,
      "eval_cer": 10.0905,
      "eval_gen_len": 14.242,
      "eval_loss": 1.8756157159805298,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 32.3794,
      "eval_samples_per_second": 61.768,
      "eval_steps_per_second": 0.247,
      "step": 3500
    },
    {
      "epoch": 0.78,
      "learning_rate": 3.696455751250272e-05,
      "loss": 2.057,
      "step": 3600
    },
    {
      "epoch": 0.8,
      "learning_rate": 3.6602159889831124e-05,
      "loss": 2.0307,
      "step": 3700
    },
    {
      "epoch": 0.83,
      "learning_rate": 3.6243386243386245e-05,
      "loss": 2.0145,
      "step": 3800
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.588098862071465e-05,
      "loss": 1.9898,
      "step": 3900
    },
    {
      "epoch": 0.87,
      "learning_rate": 3.5518590998043053e-05,
      "loss": 1.9565,
      "step": 4000
    },
    {
      "epoch": 0.87,
      "eval_cer": 9.7629,
      "eval_gen_len": 14.042,
      "eval_loss": 1.7376880645751953,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 31.0968,
      "eval_samples_per_second": 64.315,
      "eval_steps_per_second": 0.257,
      "step": 4000
    },
    {
      "epoch": 0.89,
      "learning_rate": 3.515619337537146e-05,
      "loss": 1.921,
      "step": 4100
    },
    {
      "epoch": 0.91,
      "learning_rate": 3.479379575269986e-05,
      "loss": 1.9099,
      "step": 4200
    },
    {
      "epoch": 0.93,
      "learning_rate": 3.443139813002827e-05,
      "loss": 1.8779,
      "step": 4300
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.406900050735668e-05,
      "loss": 1.8613,
      "step": 4400
    },
    {
      "epoch": 0.98,
      "learning_rate": 3.370660288468508e-05,
      "loss": 1.8319,
      "step": 4500
    },
    {
      "epoch": 0.98,
      "eval_cer": 9.4737,
      "eval_gen_len": 13.877,
      "eval_loss": 1.6244360208511353,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 31.2035,
      "eval_samples_per_second": 64.095,
      "eval_steps_per_second": 0.256,
      "step": 4500
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.3344205262013486e-05,
      "loss": 1.8114,
      "step": 4600
    },
    {
      "epoch": 1.02,
      "learning_rate": 3.298180763934189e-05,
      "loss": 1.7689,
      "step": 4700
    },
    {
      "epoch": 1.04,
      "learning_rate": 3.2619410016670295e-05,
      "loss": 1.7452,
      "step": 4800
    },
    {
      "epoch": 1.07,
      "learning_rate": 3.2257012393998696e-05,
      "loss": 1.7304,
      "step": 4900
    },
    {
      "epoch": 1.09,
      "learning_rate": 3.1894614771327104e-05,
      "loss": 1.7104,
      "step": 5000
    },
    {
      "epoch": 1.09,
      "eval_cer": 9.153,
      "eval_gen_len": 13.825,
      "eval_loss": 1.5353318452835083,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 30.755,
      "eval_samples_per_second": 65.03,
      "eval_steps_per_second": 0.26,
      "step": 5000
    },
    {
      "epoch": 1.11,
      "learning_rate": 3.1532217148655505e-05,
      "loss": 1.6878,
      "step": 5100
    },
    {
      "epoch": 1.13,
      "learning_rate": 3.116981952598391e-05,
      "loss": 1.6728,
      "step": 5200
    },
    {
      "epoch": 1.15,
      "learning_rate": 3.080742190331232e-05,
      "loss": 1.6593,
      "step": 5300
    },
    {
      "epoch": 1.17,
      "learning_rate": 3.044502428064072e-05,
      "loss": 1.6418,
      "step": 5400
    },
    {
      "epoch": 1.2,
      "learning_rate": 3.0082626657969125e-05,
      "loss": 1.6148,
      "step": 5500
    },
    {
      "epoch": 1.2,
      "eval_cer": 8.8125,
      "eval_gen_len": 13.726,
      "eval_loss": 1.4422341585159302,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 30.1297,
      "eval_samples_per_second": 66.38,
      "eval_steps_per_second": 0.266,
      "step": 5500
    },
    {
      "epoch": 1.22,
      "learning_rate": 2.972022903529753e-05,
      "loss": 1.6163,
      "step": 5600
    },
    {
      "epoch": 1.24,
      "learning_rate": 2.9357831412625937e-05,
      "loss": 1.5967,
      "step": 5700
    },
    {
      "epoch": 1.26,
      "learning_rate": 2.8995433789954342e-05,
      "loss": 1.5853,
      "step": 5800
    },
    {
      "epoch": 1.28,
      "learning_rate": 2.8633036167282746e-05,
      "loss": 1.5532,
      "step": 5900
    },
    {
      "epoch": 1.3,
      "learning_rate": 2.827063854461115e-05,
      "loss": 1.5459,
      "step": 6000
    },
    {
      "epoch": 1.3,
      "eval_cer": 8.5589,
      "eval_gen_len": 13.681,
      "eval_loss": 1.3754663467407227,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 30.0363,
      "eval_samples_per_second": 66.586,
      "eval_steps_per_second": 0.266,
      "step": 6000
    },
    {
      "epoch": 1.33,
      "learning_rate": 2.7908240921939555e-05,
      "loss": 1.5384,
      "step": 6100
    },
    {
      "epoch": 1.35,
      "learning_rate": 2.754584329926796e-05,
      "loss": 1.5169,
      "step": 6200
    },
    {
      "epoch": 1.37,
      "learning_rate": 2.7183445676596363e-05,
      "loss": 1.5045,
      "step": 6300
    },
    {
      "epoch": 1.39,
      "learning_rate": 2.6821048053924768e-05,
      "loss": 1.4945,
      "step": 6400
    },
    {
      "epoch": 1.41,
      "learning_rate": 2.6458650431253172e-05,
      "loss": 1.4837,
      "step": 6500
    },
    {
      "epoch": 1.41,
      "eval_cer": 8.2717,
      "eval_gen_len": 13.6225,
      "eval_loss": 1.3035191297531128,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 29.7288,
      "eval_samples_per_second": 67.275,
      "eval_steps_per_second": 0.269,
      "step": 6500
    },
    {
      "epoch": 1.44,
      "learning_rate": 2.609625280858158e-05,
      "loss": 1.4608,
      "step": 6600
    },
    {
      "epoch": 1.46,
      "learning_rate": 2.5733855185909984e-05,
      "loss": 1.4631,
      "step": 6700
    },
    {
      "epoch": 1.48,
      "learning_rate": 2.537145756323839e-05,
      "loss": 1.454,
      "step": 6800
    },
    {
      "epoch": 1.5,
      "learning_rate": 2.5009059940566793e-05,
      "loss": 1.4402,
      "step": 6900
    },
    {
      "epoch": 1.52,
      "learning_rate": 2.4646662317895194e-05,
      "loss": 1.4219,
      "step": 7000
    },
    {
      "epoch": 1.52,
      "eval_cer": 8.0684,
      "eval_gen_len": 13.549,
      "eval_loss": 1.2407491207122803,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 29.9599,
      "eval_samples_per_second": 66.756,
      "eval_steps_per_second": 0.267,
      "step": 7000
    },
    {
      "epoch": 1.54,
      "learning_rate": 2.4284264695223598e-05,
      "loss": 1.4081,
      "step": 7100
    },
    {
      "epoch": 1.57,
      "learning_rate": 2.3921867072552002e-05,
      "loss": 1.3969,
      "step": 7200
    },
    {
      "epoch": 1.59,
      "learning_rate": 2.355946944988041e-05,
      "loss": 1.3942,
      "step": 7300
    },
    {
      "epoch": 1.61,
      "learning_rate": 2.3197071827208815e-05,
      "loss": 1.3834,
      "step": 7400
    },
    {
      "epoch": 1.63,
      "learning_rate": 2.283467420453722e-05,
      "loss": 1.3743,
      "step": 7500
    },
    {
      "epoch": 1.63,
      "eval_cer": 7.7684,
      "eval_gen_len": 13.502,
      "eval_loss": 1.186456561088562,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 29.5504,
      "eval_samples_per_second": 67.681,
      "eval_steps_per_second": 0.271,
      "step": 7500
    },
    {
      "epoch": 1.65,
      "learning_rate": 2.2472276581865623e-05,
      "loss": 1.361,
      "step": 7600
    },
    {
      "epoch": 1.67,
      "learning_rate": 2.2109878959194028e-05,
      "loss": 1.3531,
      "step": 7700
    },
    {
      "epoch": 1.7,
      "learning_rate": 2.1747481336522432e-05,
      "loss": 1.3436,
      "step": 7800
    },
    {
      "epoch": 1.72,
      "learning_rate": 2.1385083713850836e-05,
      "loss": 1.3365,
      "step": 7900
    },
    {
      "epoch": 1.74,
      "learning_rate": 2.102268609117924e-05,
      "loss": 1.3318,
      "step": 8000
    },
    {
      "epoch": 1.74,
      "eval_cer": 7.5247,
      "eval_gen_len": 13.509,
      "eval_loss": 1.1502524614334106,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 29.2293,
      "eval_samples_per_second": 68.424,
      "eval_steps_per_second": 0.274,
      "step": 8000
    },
    {
      "epoch": 1.76,
      "learning_rate": 2.0660288468507645e-05,
      "loss": 1.3248,
      "step": 8100
    },
    {
      "epoch": 1.78,
      "learning_rate": 2.0297890845836053e-05,
      "loss": 1.312,
      "step": 8200
    },
    {
      "epoch": 1.8,
      "learning_rate": 1.9935493223164457e-05,
      "loss": 1.3043,
      "step": 8300
    },
    {
      "epoch": 1.83,
      "learning_rate": 1.957309560049286e-05,
      "loss": 1.298,
      "step": 8400
    },
    {
      "epoch": 1.85,
      "learning_rate": 1.9210697977821266e-05,
      "loss": 1.2893,
      "step": 8500
    },
    {
      "epoch": 1.85,
      "eval_cer": 7.3826,
      "eval_gen_len": 13.456,
      "eval_loss": 1.108539342880249,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 29.5564,
      "eval_samples_per_second": 67.667,
      "eval_steps_per_second": 0.271,
      "step": 8500
    },
    {
      "epoch": 1.87,
      "learning_rate": 3.1308980213089806e-05,
      "loss": 1.2603,
      "step": 8600
    },
    {
      "epoch": 1.89,
      "learning_rate": 3.109154163948685e-05,
      "loss": 1.2545,
      "step": 8700
    },
    {
      "epoch": 1.91,
      "learning_rate": 3.0874103065883894e-05,
      "loss": 1.2411,
      "step": 8800
    },
    {
      "epoch": 1.94,
      "learning_rate": 3.0656664492280935e-05,
      "loss": 1.2354,
      "step": 8900
    },
    {
      "epoch": 1.96,
      "learning_rate": 3.043922591867798e-05,
      "loss": 1.2243,
      "step": 9000
    },
    {
      "epoch": 1.98,
      "learning_rate": 3.022178734507502e-05,
      "loss": 1.2228,
      "step": 9100
    },
    {
      "epoch": 2.0,
      "eval_cer": 7.0411,
      "eval_gen_len": 13.3935,
      "eval_loss": 1.0505541563034058,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 29.0334,
      "eval_samples_per_second": 68.886,
      "eval_steps_per_second": 0.276,
      "step": 9198
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.0004348771472063e-05,
      "loss": 1.2108,
      "step": 9200
    },
    {
      "epoch": 2.02,
      "learning_rate": 2.9786910197869107e-05,
      "loss": 1.1899,
      "step": 9300
    },
    {
      "epoch": 2.04,
      "learning_rate": 2.9569471624266148e-05,
      "loss": 1.1901,
      "step": 9400
    },
    {
      "epoch": 2.07,
      "learning_rate": 2.9352033050663192e-05,
      "loss": 1.1772,
      "step": 9500
    },
    {
      "epoch": 2.09,
      "learning_rate": 2.9134594477060236e-05,
      "loss": 1.1657,
      "step": 9600
    },
    {
      "epoch": 2.11,
      "learning_rate": 2.8917155903457276e-05,
      "loss": 1.1618,
      "step": 9700
    },
    {
      "epoch": 2.13,
      "learning_rate": 2.869971732985432e-05,
      "loss": 1.1517,
      "step": 9800
    },
    {
      "epoch": 2.15,
      "learning_rate": 2.8482278756251364e-05,
      "loss": 1.1428,
      "step": 9900
    },
    {
      "epoch": 2.17,
      "learning_rate": 2.826701456838443e-05,
      "loss": 1.1386,
      "step": 10000
    },
    {
      "epoch": 2.2,
      "learning_rate": 2.8049575994781475e-05,
      "loss": 1.1285,
      "step": 10100
    },
    {
      "epoch": 2.22,
      "learning_rate": 2.7832137421178516e-05,
      "loss": 1.1213,
      "step": 10200
    },
    {
      "epoch": 2.24,
      "learning_rate": 2.761469884757556e-05,
      "loss": 1.1155,
      "step": 10300
    },
    {
      "epoch": 2.26,
      "learning_rate": 2.7397260273972603e-05,
      "loss": 1.1163,
      "step": 10400
    },
    {
      "epoch": 2.28,
      "learning_rate": 2.7179821700369644e-05,
      "loss": 1.1055,
      "step": 10500
    },
    {
      "epoch": 2.3,
      "learning_rate": 2.6962383126766688e-05,
      "loss": 1.0981,
      "step": 10600
    },
    {
      "epoch": 2.33,
      "learning_rate": 2.6744944553163732e-05,
      "loss": 1.0893,
      "step": 10700
    },
    {
      "epoch": 2.35,
      "learning_rate": 2.6527505979560773e-05,
      "loss": 1.0868,
      "step": 10800
    },
    {
      "epoch": 2.37,
      "learning_rate": 2.6310067405957817e-05,
      "loss": 1.0793,
      "step": 10900
    },
    {
      "epoch": 2.39,
      "learning_rate": 2.609262883235486e-05,
      "loss": 1.0699,
      "step": 11000
    },
    {
      "epoch": 2.41,
      "learning_rate": 2.58751902587519e-05,
      "loss": 1.0628,
      "step": 11100
    },
    {
      "epoch": 2.44,
      "learning_rate": 2.5657751685148945e-05,
      "loss": 1.0506,
      "step": 11200
    },
    {
      "epoch": 2.46,
      "learning_rate": 2.544031311154599e-05,
      "loss": 1.0437,
      "step": 11300
    },
    {
      "epoch": 2.48,
      "learning_rate": 2.522287453794303e-05,
      "loss": 1.0397,
      "step": 11400
    },
    {
      "epoch": 2.5,
      "learning_rate": 2.5005435964340073e-05,
      "loss": 1.0436,
      "step": 11500
    },
    {
      "epoch": 2.52,
      "learning_rate": 2.4787997390737117e-05,
      "loss": 1.0313,
      "step": 11600
    },
    {
      "epoch": 2.54,
      "learning_rate": 2.4570558817134158e-05,
      "loss": 1.0209,
      "step": 11700
    },
    {
      "epoch": 2.57,
      "learning_rate": 2.4353120243531202e-05,
      "loss": 1.0199,
      "step": 11800
    },
    {
      "epoch": 2.59,
      "learning_rate": 2.4135681669928246e-05,
      "loss": 1.0153,
      "step": 11900
    },
    {
      "epoch": 2.61,
      "learning_rate": 2.3918243096325287e-05,
      "loss": 1.0073,
      "step": 12000
    },
    {
      "epoch": 2.63,
      "learning_rate": 2.370080452272233e-05,
      "loss": 1.0046,
      "step": 12100
    },
    {
      "epoch": 2.65,
      "learning_rate": 2.3483365949119374e-05,
      "loss": 0.9931,
      "step": 12200
    },
    {
      "epoch": 2.67,
      "learning_rate": 2.3265927375516415e-05,
      "loss": 0.9957,
      "step": 12300
    },
    {
      "epoch": 2.7,
      "learning_rate": 2.304848880191346e-05,
      "loss": 0.9856,
      "step": 12400
    },
    {
      "epoch": 2.72,
      "learning_rate": 2.2831050228310503e-05,
      "loss": 0.9821,
      "step": 12500
    },
    {
      "epoch": 2.74,
      "learning_rate": 2.2613611654707543e-05,
      "loss": 0.9727,
      "step": 12600
    },
    {
      "epoch": 2.76,
      "learning_rate": 2.2396173081104587e-05,
      "loss": 0.9708,
      "step": 12700
    },
    {
      "epoch": 2.78,
      "learning_rate": 2.217873450750163e-05,
      "loss": 0.9737,
      "step": 12800
    },
    {
      "epoch": 2.8,
      "learning_rate": 2.1961295933898672e-05,
      "loss": 0.969,
      "step": 12900
    },
    {
      "epoch": 2.83,
      "learning_rate": 2.1743857360295716e-05,
      "loss": 0.9544,
      "step": 13000
    },
    {
      "epoch": 2.85,
      "learning_rate": 2.152641878669276e-05,
      "loss": 0.9515,
      "step": 13100
    },
    {
      "epoch": 2.87,
      "learning_rate": 2.1311154598825833e-05,
      "loss": 0.9502,
      "step": 13200
    },
    {
      "epoch": 2.89,
      "learning_rate": 2.1093716025222877e-05,
      "loss": 0.9409,
      "step": 13300
    },
    {
      "epoch": 2.91,
      "learning_rate": 2.0876277451619918e-05,
      "loss": 0.9451,
      "step": 13400
    },
    {
      "epoch": 2.94,
      "learning_rate": 2.0658838878016962e-05,
      "loss": 0.9388,
      "step": 13500
    },
    {
      "epoch": 2.96,
      "learning_rate": 2.0441400304414006e-05,
      "loss": 0.9337,
      "step": 13600
    },
    {
      "epoch": 2.98,
      "learning_rate": 2.0223961730811046e-05,
      "loss": 0.9343,
      "step": 13700
    },
    {
      "epoch": 3.0,
      "eval_cer": 5.5303,
      "eval_gen_len": 13.2935,
      "eval_loss": 0.776914656162262,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 28.2855,
      "eval_samples_per_second": 70.708,
      "eval_steps_per_second": 0.283,
      "step": 13797
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.000652315720809e-05,
      "loss": 0.922,
      "step": 13800
    },
    {
      "epoch": 3.02,
      "learning_rate": 1.9789084583605134e-05,
      "loss": 0.8943,
      "step": 13900
    },
    {
      "epoch": 3.04,
      "learning_rate": 1.9571646010002175e-05,
      "loss": 0.8884,
      "step": 14000
    },
    {
      "epoch": 3.07,
      "learning_rate": 1.935420743639922e-05,
      "loss": 0.8858,
      "step": 14100
    },
    {
      "epoch": 3.09,
      "learning_rate": 1.9136768862796263e-05,
      "loss": 0.8831,
      "step": 14200
    },
    {
      "epoch": 3.11,
      "learning_rate": 1.8919330289193303e-05,
      "loss": 0.8857,
      "step": 14300
    },
    {
      "epoch": 3.13,
      "learning_rate": 1.8701891715590347e-05,
      "loss": 0.8825,
      "step": 14400
    },
    {
      "epoch": 3.15,
      "learning_rate": 1.848445314198739e-05,
      "loss": 0.8758,
      "step": 14500
    },
    {
      "epoch": 3.17,
      "learning_rate": 1.8267014568384432e-05,
      "loss": 0.8742,
      "step": 14600
    },
    {
      "epoch": 3.2,
      "learning_rate": 1.8049575994781476e-05,
      "loss": 0.8762,
      "step": 14700
    },
    {
      "epoch": 3.22,
      "learning_rate": 1.783213742117852e-05,
      "loss": 0.8654,
      "step": 14800
    },
    {
      "epoch": 3.24,
      "learning_rate": 1.761469884757556e-05,
      "loss": 0.8697,
      "step": 14900
    },
    {
      "epoch": 3.26,
      "learning_rate": 1.7397260273972604e-05,
      "loss": 0.8648,
      "step": 15000
    },
    {
      "epoch": 3.28,
      "learning_rate": 1.717982170036965e-05,
      "loss": 0.859,
      "step": 15100
    },
    {
      "epoch": 3.31,
      "learning_rate": 1.696238312676669e-05,
      "loss": 0.8616,
      "step": 15200
    },
    {
      "epoch": 3.33,
      "learning_rate": 1.6744944553163733e-05,
      "loss": 0.8519,
      "step": 15300
    },
    {
      "epoch": 3.35,
      "learning_rate": 1.6527505979560777e-05,
      "loss": 0.8493,
      "step": 15400
    },
    {
      "epoch": 3.37,
      "learning_rate": 1.6310067405957817e-05,
      "loss": 0.8478,
      "step": 15500
    },
    {
      "epoch": 3.39,
      "learning_rate": 1.609262883235486e-05,
      "loss": 0.8514,
      "step": 15600
    },
    {
      "epoch": 3.41,
      "learning_rate": 1.5875190258751905e-05,
      "loss": 0.8435,
      "step": 15700
    },
    {
      "epoch": 3.44,
      "learning_rate": 1.5657751685148946e-05,
      "loss": 0.8436,
      "step": 15800
    },
    {
      "epoch": 3.46,
      "learning_rate": 1.544031311154599e-05,
      "loss": 0.8393,
      "step": 15900
    },
    {
      "epoch": 3.48,
      "learning_rate": 1.5222874537943032e-05,
      "loss": 0.8355,
      "step": 16000
    },
    {
      "epoch": 3.5,
      "learning_rate": 1.5005435964340076e-05,
      "loss": 0.8364,
      "step": 16100
    },
    {
      "epoch": 3.52,
      "learning_rate": 1.4787997390737118e-05,
      "loss": 0.8331,
      "step": 16200
    },
    {
      "epoch": 3.54,
      "learning_rate": 1.457055881713416e-05,
      "loss": 0.829,
      "step": 16300
    },
    {
      "epoch": 3.57,
      "learning_rate": 1.4353120243531205e-05,
      "loss": 0.8273,
      "step": 16400
    },
    {
      "epoch": 3.59,
      "learning_rate": 1.4135681669928247e-05,
      "loss": 0.8261,
      "step": 16500
    },
    {
      "epoch": 3.61,
      "learning_rate": 1.3918243096325289e-05,
      "loss": 0.8252,
      "step": 16600
    },
    {
      "epoch": 3.63,
      "learning_rate": 1.3700804522722333e-05,
      "loss": 0.8269,
      "step": 16700
    },
    {
      "epoch": 3.65,
      "learning_rate": 1.3483365949119375e-05,
      "loss": 0.8206,
      "step": 16800
    },
    {
      "epoch": 3.67,
      "learning_rate": 1.3265927375516418e-05,
      "loss": 0.8176,
      "step": 16900
    },
    {
      "epoch": 3.7,
      "learning_rate": 1.3048488801913462e-05,
      "loss": 0.8119,
      "step": 17000
    },
    {
      "epoch": 3.72,
      "learning_rate": 1.2831050228310504e-05,
      "loss": 0.813,
      "step": 17100
    },
    {
      "epoch": 3.74,
      "learning_rate": 1.2615786040443576e-05,
      "loss": 0.8097,
      "step": 17200
    },
    {
      "epoch": 3.76,
      "learning_rate": 1.2398347466840618e-05,
      "loss": 0.8073,
      "step": 17300
    },
    {
      "epoch": 3.78,
      "learning_rate": 1.218090889323766e-05,
      "loss": 0.8016,
      "step": 17400
    },
    {
      "epoch": 3.81,
      "learning_rate": 1.1963470319634704e-05,
      "loss": 0.8068,
      "step": 17500
    },
    {
      "epoch": 3.83,
      "learning_rate": 1.1746031746031746e-05,
      "loss": 0.8034,
      "step": 17600
    },
    {
      "epoch": 3.85,
      "learning_rate": 1.1528593172428789e-05,
      "loss": 0.8051,
      "step": 17700
    },
    {
      "epoch": 3.87,
      "learning_rate": 1.1311154598825833e-05,
      "loss": 0.8043,
      "step": 17800
    },
    {
      "epoch": 3.89,
      "learning_rate": 1.1095890410958904e-05,
      "loss": 0.801,
      "step": 17900
    },
    {
      "epoch": 3.91,
      "learning_rate": 1.0878451837355948e-05,
      "loss": 0.7976,
      "step": 18000
    },
    {
      "epoch": 3.94,
      "learning_rate": 1.066101326375299e-05,
      "loss": 0.7968,
      "step": 18100
    },
    {
      "epoch": 3.96,
      "learning_rate": 1.0443574690150033e-05,
      "loss": 0.7983,
      "step": 18200
    },
    {
      "epoch": 3.98,
      "learning_rate": 1.0226136116547077e-05,
      "loss": 0.7915,
      "step": 18300
    },
    {
      "epoch": 4.0,
      "eval_cer": 4.8928,
      "eval_gen_len": 13.209,
      "eval_loss": 0.666305661201477,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 28.2589,
      "eval_samples_per_second": 70.774,
      "eval_steps_per_second": 0.283,
      "step": 18396
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.000869754294412e-05,
      "loss": 0.7887,
      "step": 18400
    },
    {
      "epoch": 4.02,
      "learning_rate": 9.791258969341161e-06,
      "loss": 0.7713,
      "step": 18500
    },
    {
      "epoch": 4.04,
      "learning_rate": 9.573820395738205e-06,
      "loss": 0.7674,
      "step": 18600
    },
    {
      "epoch": 4.07,
      "learning_rate": 9.356381822135248e-06,
      "loss": 0.7697,
      "step": 18700
    },
    {
      "epoch": 4.09,
      "learning_rate": 9.13894324853229e-06,
      "loss": 0.7688,
      "step": 18800
    },
    {
      "epoch": 4.11,
      "learning_rate": 8.921504674929334e-06,
      "loss": 0.7681,
      "step": 18900
    },
    {
      "epoch": 4.13,
      "learning_rate": 8.704066101326376e-06,
      "loss": 0.7649,
      "step": 19000
    },
    {
      "epoch": 4.15,
      "learning_rate": 8.486627527723418e-06,
      "loss": 0.7664,
      "step": 19100
    },
    {
      "epoch": 4.17,
      "learning_rate": 8.269188954120462e-06,
      "loss": 0.7619,
      "step": 19200
    },
    {
      "epoch": 4.2,
      "learning_rate": 8.051750380517505e-06,
      "loss": 0.7658,
      "step": 19300
    },
    {
      "epoch": 4.22,
      "learning_rate": 7.834311806914547e-06,
      "loss": 0.7574,
      "step": 19400
    },
    {
      "epoch": 4.24,
      "learning_rate": 7.61687323331159e-06,
      "loss": 0.7614,
      "step": 19500
    },
    {
      "epoch": 4.26,
      "learning_rate": 7.399434659708633e-06,
      "loss": 0.7564,
      "step": 19600
    },
    {
      "epoch": 4.28,
      "learning_rate": 7.181996086105676e-06,
      "loss": 0.7611,
      "step": 19700
    },
    {
      "epoch": 4.31,
      "learning_rate": 6.9645575125027185e-06,
      "loss": 0.7582,
      "step": 19800
    },
    {
      "epoch": 4.33,
      "learning_rate": 6.747118938899762e-06,
      "loss": 0.7558,
      "step": 19900
    },
    {
      "epoch": 4.35,
      "learning_rate": 6.529680365296805e-06,
      "loss": 0.76,
      "step": 20000
    },
    {
      "epoch": 4.37,
      "learning_rate": 6.312241791693847e-06,
      "loss": 0.7605,
      "step": 20100
    },
    {
      "epoch": 4.39,
      "learning_rate": 6.094803218090889e-06,
      "loss": 0.7516,
      "step": 20200
    },
    {
      "epoch": 4.41,
      "learning_rate": 5.877364644487932e-06,
      "loss": 0.7506,
      "step": 20300
    },
    {
      "epoch": 4.44,
      "learning_rate": 5.6599260708849755e-06,
      "loss": 0.747,
      "step": 20400
    },
    {
      "epoch": 4.46,
      "learning_rate": 5.442487497282018e-06,
      "loss": 0.7512,
      "step": 20500
    },
    {
      "epoch": 4.48,
      "learning_rate": 5.225048923679061e-06,
      "loss": 0.751,
      "step": 20600
    },
    {
      "epoch": 4.5,
      "learning_rate": 5.007610350076104e-06,
      "loss": 0.7515,
      "step": 20700
    },
    {
      "epoch": 4.52,
      "learning_rate": 4.790171776473146e-06,
      "loss": 0.7535,
      "step": 20800
    },
    {
      "epoch": 4.54,
      "learning_rate": 4.572733202870189e-06,
      "loss": 0.7511,
      "step": 20900
    },
    {
      "epoch": 4.57,
      "learning_rate": 4.3552946292672325e-06,
      "loss": 0.7526,
      "step": 21000
    },
    {
      "epoch": 4.59,
      "learning_rate": 4.137856055664275e-06,
      "loss": 0.7416,
      "step": 21100
    },
    {
      "epoch": 4.61,
      "learning_rate": 3.920417482061318e-06,
      "loss": 0.7506,
      "step": 21200
    },
    {
      "epoch": 4.63,
      "learning_rate": 3.7029789084583605e-06,
      "loss": 0.7478,
      "step": 21300
    },
    {
      "epoch": 4.65,
      "learning_rate": 3.4855403348554037e-06,
      "loss": 0.7454,
      "step": 21400
    },
    {
      "epoch": 4.67,
      "learning_rate": 3.2681017612524463e-06,
      "loss": 0.7512,
      "step": 21500
    },
    {
      "epoch": 4.7,
      "learning_rate": 3.050663187649489e-06,
      "loss": 0.747,
      "step": 21600
    },
    {
      "epoch": 4.72,
      "learning_rate": 2.833224614046532e-06,
      "loss": 0.7423,
      "step": 21700
    },
    {
      "epoch": 4.74,
      "learning_rate": 2.615786040443575e-06,
      "loss": 0.7408,
      "step": 21800
    },
    {
      "epoch": 4.76,
      "learning_rate": 2.3983474668406175e-06,
      "loss": 0.7457,
      "step": 21900
    },
    {
      "epoch": 4.78,
      "learning_rate": 2.1809088932376606e-06,
      "loss": 0.7447,
      "step": 22000
    },
    {
      "epoch": 4.81,
      "learning_rate": 1.9634703196347033e-06,
      "loss": 0.7346,
      "step": 22100
    },
    {
      "epoch": 4.83,
      "learning_rate": 1.7460317460317462e-06,
      "loss": 0.7404,
      "step": 22200
    },
    {
      "epoch": 4.85,
      "learning_rate": 1.528593172428789e-06,
      "loss": 0.7494,
      "step": 22300
    },
    {
      "epoch": 4.87,
      "learning_rate": 1.3111545988258318e-06,
      "loss": 0.7361,
      "step": 22400
    },
    {
      "epoch": 4.89,
      "learning_rate": 1.0937160252228747e-06,
      "loss": 0.7417,
      "step": 22500
    },
    {
      "epoch": 4.91,
      "learning_rate": 8.762774516199173e-07,
      "loss": 0.7437,
      "step": 22600
    },
    {
      "epoch": 4.94,
      "learning_rate": 6.588388780169603e-07,
      "loss": 0.7395,
      "step": 22700
    },
    {
      "epoch": 4.96,
      "learning_rate": 4.414003044140031e-07,
      "loss": 0.7393,
      "step": 22800
    },
    {
      "epoch": 4.98,
      "learning_rate": 2.2396173081104588e-07,
      "loss": 0.7436,
      "step": 22900
    },
    {
      "epoch": 5.0,
      "eval_cer": 4.7349,
      "eval_gen_len": 13.2035,
      "eval_loss": 0.6352977752685547,
      "eval_rouge1": 0.0,
      "eval_rouge2": 0.0,
      "eval_rougeL": 0.0,
      "eval_rougeLsum": 0.0,
      "eval_runtime": 27.4711,
      "eval_samples_per_second": 72.804,
      "eval_steps_per_second": 0.291,
      "step": 22995
    },
    {
      "epoch": 5.0,
      "step": 22995,
      "total_flos": 1.777894335479808e+17,
      "train_loss": 0.564045099207411,
      "train_runtime": 5933.1558,
      "train_samples_per_second": 1239.969,
      "train_steps_per_second": 3.876
    }
  ],
  "logging_steps": 100,
  "max_steps": 22995,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 1.777894335479808e+17,
  "trial_name": null,
  "trial_params": null
}