{
  "best_metric": 22.042796384901646,
  "best_model_checkpoint": "BA_Model_V2/checkpoint-3000",
  "epoch": 17.985611510791365,
  "eval_steps": 1000,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.36,
      "grad_norm": 8.64929485321045,
      "learning_rate": 1.9600000000000003e-06,
      "loss": 1.2156,
      "step": 100
    },
    {
      "epoch": 0.72,
      "grad_norm": 3.87185001373291,
      "learning_rate": 3.96e-06,
      "loss": 0.3891,
      "step": 200
    },
    {
      "epoch": 1.08,
      "grad_norm": 3.1216392517089844,
      "learning_rate": 5.9600000000000005e-06,
      "loss": 0.2699,
      "step": 300
    },
    {
      "epoch": 1.44,
      "grad_norm": 6.549575328826904,
      "learning_rate": 7.960000000000002e-06,
      "loss": 0.2151,
      "step": 400
    },
    {
      "epoch": 1.8,
      "grad_norm": 2.978250026702881,
      "learning_rate": 9.960000000000001e-06,
      "loss": 0.2213,
      "step": 500
    },
    {
      "epoch": 2.16,
      "grad_norm": 8.912956237792969,
      "learning_rate": 9.782222222222222e-06,
      "loss": 0.1881,
      "step": 600
    },
    {
      "epoch": 2.52,
      "grad_norm": 2.4266467094421387,
      "learning_rate": 9.56e-06,
      "loss": 0.1264,
      "step": 700
    },
    {
      "epoch": 2.88,
      "grad_norm": 2.5170629024505615,
      "learning_rate": 9.33777777777778e-06,
      "loss": 0.1319,
      "step": 800
    },
    {
      "epoch": 3.24,
      "grad_norm": 1.9321565628051758,
      "learning_rate": 9.115555555555556e-06,
      "loss": 0.0872,
      "step": 900
    },
    {
      "epoch": 3.6,
      "grad_norm": 2.928337335586548,
      "learning_rate": 8.893333333333333e-06,
      "loss": 0.0687,
      "step": 1000
    },
    {
      "epoch": 3.6,
      "eval_cer": 12.433448002116961,
      "eval_loss": 0.37324240803718567,
      "eval_runtime": 946.036,
      "eval_samples_per_second": 1.568,
      "eval_steps_per_second": 0.197,
      "eval_wer": 22.268740031897927,
      "step": 1000
    },
    {
      "epoch": 3.96,
      "grad_norm": 2.3376998901367188,
      "learning_rate": 8.671111111111113e-06,
      "loss": 0.0724,
      "step": 1100
    },
    {
      "epoch": 4.32,
      "grad_norm": 1.9221140146255493,
      "learning_rate": 8.448888888888889e-06,
      "loss": 0.0423,
      "step": 1200
    },
    {
      "epoch": 4.68,
      "grad_norm": 1.7310447692871094,
      "learning_rate": 8.226666666666667e-06,
      "loss": 0.0423,
      "step": 1300
    },
    {
      "epoch": 5.04,
      "grad_norm": 1.6635299921035767,
      "learning_rate": 8.004444444444445e-06,
      "loss": 0.0425,
      "step": 1400
    },
    {
      "epoch": 5.4,
      "grad_norm": 1.2995408773422241,
      "learning_rate": 7.782222222222223e-06,
      "loss": 0.0272,
      "step": 1500
    },
    {
      "epoch": 5.76,
      "grad_norm": 2.0113301277160645,
      "learning_rate": 7.5600000000000005e-06,
      "loss": 0.0307,
      "step": 1600
    },
    {
      "epoch": 6.12,
      "grad_norm": 1.9505223035812378,
      "learning_rate": 7.337777777777778e-06,
      "loss": 0.0244,
      "step": 1700
    },
    {
      "epoch": 6.47,
      "grad_norm": 1.1860969066619873,
      "learning_rate": 7.115555555555557e-06,
      "loss": 0.018,
      "step": 1800
    },
    {
      "epoch": 6.83,
      "grad_norm": 0.8492623567581177,
      "learning_rate": 6.893333333333334e-06,
      "loss": 0.0196,
      "step": 1900
    },
    {
      "epoch": 7.19,
      "grad_norm": 1.2874751091003418,
      "learning_rate": 6.671111111111112e-06,
      "loss": 0.0151,
      "step": 2000
    },
    {
      "epoch": 7.19,
      "eval_cer": 12.361471288700715,
      "eval_loss": 0.4464028477668762,
      "eval_runtime": 940.6773,
      "eval_samples_per_second": 1.577,
      "eval_steps_per_second": 0.198,
      "eval_wer": 22.610978203083466,
      "step": 2000
    },
    {
      "epoch": 7.55,
      "grad_norm": 0.9422910809516907,
      "learning_rate": 6.448888888888889e-06,
      "loss": 0.0117,
      "step": 2100
    },
    {
      "epoch": 7.91,
      "grad_norm": 0.8238167762756348,
      "learning_rate": 6.2266666666666675e-06,
      "loss": 0.0119,
      "step": 2200
    },
    {
      "epoch": 8.27,
      "grad_norm": 0.7173901796340942,
      "learning_rate": 6.004444444444445e-06,
      "loss": 0.0094,
      "step": 2300
    },
    {
      "epoch": 8.63,
      "grad_norm": 0.7112721800804138,
      "learning_rate": 5.782222222222222e-06,
      "loss": 0.0103,
      "step": 2400
    },
    {
      "epoch": 8.99,
      "grad_norm": 0.7713698148727417,
      "learning_rate": 5.560000000000001e-06,
      "loss": 0.0087,
      "step": 2500
    },
    {
      "epoch": 9.35,
      "grad_norm": 0.798090934753418,
      "learning_rate": 5.337777777777779e-06,
      "loss": 0.0045,
      "step": 2600
    },
    {
      "epoch": 9.71,
      "grad_norm": 0.2931395471096039,
      "learning_rate": 5.115555555555556e-06,
      "loss": 0.0062,
      "step": 2700
    },
    {
      "epoch": 10.07,
      "grad_norm": 0.447310209274292,
      "learning_rate": 4.893333333333334e-06,
      "loss": 0.0044,
      "step": 2800
    },
    {
      "epoch": 10.43,
      "grad_norm": 0.07779108732938766,
      "learning_rate": 4.6711111111111115e-06,
      "loss": 0.0045,
      "step": 2900
    },
    {
      "epoch": 10.79,
      "grad_norm": 0.874796986579895,
      "learning_rate": 4.448888888888889e-06,
      "loss": 0.0039,
      "step": 3000
    },
    {
      "epoch": 10.79,
      "eval_cer": 12.351944958983859,
      "eval_loss": 0.4558783173561096,
      "eval_runtime": 949.821,
      "eval_samples_per_second": 1.561,
      "eval_steps_per_second": 0.196,
      "eval_wer": 22.042796384901646,
      "step": 3000
    },
    {
      "epoch": 11.15,
      "grad_norm": 0.058852240443229675,
      "learning_rate": 4.226666666666667e-06,
      "loss": 0.0029,
      "step": 3100
    },
    {
      "epoch": 11.51,
      "grad_norm": 0.043637294322252274,
      "learning_rate": 4.004444444444445e-06,
      "loss": 0.002,
      "step": 3200
    },
    {
      "epoch": 11.87,
      "grad_norm": 0.03788485378026962,
      "learning_rate": 3.782222222222223e-06,
      "loss": 0.0028,
      "step": 3300
    },
    {
      "epoch": 12.23,
      "grad_norm": 0.0388493537902832,
      "learning_rate": 3.5600000000000002e-06,
      "loss": 0.0021,
      "step": 3400
    },
    {
      "epoch": 12.59,
      "grad_norm": 0.36381232738494873,
      "learning_rate": 3.337777777777778e-06,
      "loss": 0.0014,
      "step": 3500
    },
    {
      "epoch": 12.95,
      "grad_norm": 0.0187509935349226,
      "learning_rate": 3.1155555555555555e-06,
      "loss": 0.0015,
      "step": 3600
    },
    {
      "epoch": 13.31,
      "grad_norm": 0.017583340406417847,
      "learning_rate": 2.8933333333333337e-06,
      "loss": 0.0016,
      "step": 3700
    },
    {
      "epoch": 13.67,
      "grad_norm": 0.019498176872730255,
      "learning_rate": 2.6711111111111116e-06,
      "loss": 0.0012,
      "step": 3800
    },
    {
      "epoch": 14.03,
      "grad_norm": 0.019207799807190895,
      "learning_rate": 2.448888888888889e-06,
      "loss": 0.0008,
      "step": 3900
    },
    {
      "epoch": 14.39,
      "grad_norm": 0.016761431470513344,
      "learning_rate": 2.226666666666667e-06,
      "loss": 0.0008,
      "step": 4000
    },
    {
      "epoch": 14.39,
      "eval_cer": 12.13548557819529,
      "eval_loss": 0.5066311359405518,
      "eval_runtime": 938.9637,
      "eval_samples_per_second": 1.579,
      "eval_steps_per_second": 0.198,
      "eval_wer": 22.155768208399788,
      "step": 4000
    },
    {
      "epoch": 14.75,
      "grad_norm": 0.014156104065477848,
      "learning_rate": 2.0044444444444446e-06,
      "loss": 0.0009,
      "step": 4100
    },
    {
      "epoch": 15.11,
      "grad_norm": 0.01545019168406725,
      "learning_rate": 1.7822222222222225e-06,
      "loss": 0.0008,
      "step": 4200
    },
    {
      "epoch": 15.47,
      "grad_norm": 0.011677294038236141,
      "learning_rate": 1.56e-06,
      "loss": 0.0006,
      "step": 4300
    },
    {
      "epoch": 15.83,
      "grad_norm": 0.018884949386119843,
      "learning_rate": 1.337777777777778e-06,
      "loss": 0.0006,
      "step": 4400
    },
    {
      "epoch": 16.19,
      "grad_norm": 0.013283302076160908,
      "learning_rate": 1.1155555555555558e-06,
      "loss": 0.0007,
      "step": 4500
    },
    {
      "epoch": 16.55,
      "grad_norm": 0.016208523884415627,
      "learning_rate": 8.933333333333334e-07,
      "loss": 0.0007,
      "step": 4600
    },
    {
      "epoch": 16.91,
      "grad_norm": 0.017564095556735992,
      "learning_rate": 6.711111111111111e-07,
      "loss": 0.0005,
      "step": 4700
    },
    {
      "epoch": 17.27,
      "grad_norm": 0.011613059788942337,
      "learning_rate": 4.488888888888889e-07,
      "loss": 0.0005,
      "step": 4800
    },
    {
      "epoch": 17.63,
      "grad_norm": 0.014883192256093025,
      "learning_rate": 2.266666666666667e-07,
      "loss": 0.0006,
      "step": 4900
    },
    {
      "epoch": 17.99,
      "grad_norm": 0.009823828004300594,
      "learning_rate": 4.444444444444445e-09,
      "loss": 0.0005,
      "step": 5000
    },
    {
      "epoch": 17.99,
      "eval_cer": 12.508070918232336,
      "eval_loss": 0.5289328098297119,
      "eval_runtime": 942.463,
      "eval_samples_per_second": 1.574,
      "eval_steps_per_second": 0.197,
      "eval_wer": 22.760499734183945,
      "step": 5000
    }
  ],
  "logging_steps": 100,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 18,
  "save_steps": 1000,
  "total_flos": 1.696010243125248e+20,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}