{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 100.0,
  "global_step": 5400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 3.7,
      "learning_rate": 5.909999999999999e-05,
      "loss": 11.4212,
      "step": 200
    },
    {
      "epoch": 7.41,
      "learning_rate": 0.0001191,
      "loss": 3.3931,
      "step": 400
    },
    {
      "epoch": 9.26,
      "eval_loss": 2.6012766361236572,
      "eval_runtime": 13.0763,
      "eval_samples_per_second": 27.531,
      "eval_steps_per_second": 3.441,
      "eval_wer": 0.991375931007448,
      "step": 500
    },
    {
      "epoch": 11.11,
      "learning_rate": 0.0001791,
      "loss": 2.2716,
      "step": 600
    },
    {
      "epoch": 14.81,
      "learning_rate": 0.00023909999999999998,
      "loss": 0.5568,
      "step": 800
    },
    {
      "epoch": 18.52,
      "learning_rate": 0.0002991,
      "loss": 0.3457,
      "step": 1000
    },
    {
      "epoch": 18.52,
      "eval_loss": 1.0315884351730347,
      "eval_runtime": 12.6921,
      "eval_samples_per_second": 28.364,
      "eval_steps_per_second": 3.546,
      "eval_wer": 0.6452371618972952,
      "step": 1000
    },
    {
      "epoch": 22.22,
      "learning_rate": 0.00032255581395348835,
      "loss": 0.2708,
      "step": 1200
    },
    {
      "epoch": 25.93,
      "learning_rate": 0.00030720697674418605,
      "loss": 0.2118,
      "step": 1400
    },
    {
      "epoch": 27.78,
      "eval_loss": 1.1182368993759155,
      "eval_runtime": 12.9665,
      "eval_samples_per_second": 27.764,
      "eval_steps_per_second": 3.47,
      "eval_wer": 0.6354370834966679,
      "step": 1500
    },
    {
      "epoch": 29.63,
      "learning_rate": 0.00029185813953488374,
      "loss": 0.1765,
      "step": 1600
    },
    {
      "epoch": 33.33,
      "learning_rate": 0.0002765093023255814,
      "loss": 0.1404,
      "step": 1800
    },
    {
      "epoch": 37.04,
      "learning_rate": 0.00026116046511627907,
      "loss": 0.1245,
      "step": 2000
    },
    {
      "epoch": 37.04,
      "eval_loss": 1.2044304609298706,
      "eval_runtime": 12.7752,
      "eval_samples_per_second": 28.18,
      "eval_steps_per_second": 3.522,
      "eval_wer": 0.6295570364562917,
      "step": 2000
    },
    {
      "epoch": 40.74,
      "learning_rate": 0.00024581162790697676,
      "loss": 0.1166,
      "step": 2200
    },
    {
      "epoch": 44.44,
      "learning_rate": 0.00023046279069767443,
      "loss": 0.1068,
      "step": 2400
    },
    {
      "epoch": 46.3,
      "eval_loss": 1.1400295495986938,
      "eval_runtime": 12.7392,
      "eval_samples_per_second": 28.259,
      "eval_steps_per_second": 3.532,
      "eval_wer": 0.582124656997256,
      "step": 2500
    },
    {
      "epoch": 48.15,
      "learning_rate": 0.0002151139534883721,
      "loss": 0.0917,
      "step": 2600
    },
    {
      "epoch": 51.85,
      "learning_rate": 0.00019976511627906976,
      "loss": 0.0842,
      "step": 2800
    },
    {
      "epoch": 55.56,
      "learning_rate": 0.00018441627906976745,
      "loss": 0.0806,
      "step": 3000
    },
    {
      "epoch": 55.56,
      "eval_loss": 1.2010185718536377,
      "eval_runtime": 12.8794,
      "eval_samples_per_second": 27.952,
      "eval_steps_per_second": 3.494,
      "eval_wer": 0.5785966287730302,
      "step": 3000
    },
    {
      "epoch": 59.26,
      "learning_rate": 0.00016906744186046509,
      "loss": 0.0719,
      "step": 3200
    },
    {
      "epoch": 62.96,
      "learning_rate": 0.00015371860465116278,
      "loss": 0.0679,
      "step": 3400
    },
    {
      "epoch": 64.81,
      "eval_loss": 1.1644504070281982,
      "eval_runtime": 12.9878,
      "eval_samples_per_second": 27.718,
      "eval_steps_per_second": 3.465,
      "eval_wer": 0.546452371618973,
      "step": 3500
    },
    {
      "epoch": 66.67,
      "learning_rate": 0.00013836976744186047,
      "loss": 0.0582,
      "step": 3600
    },
    {
      "epoch": 70.37,
      "learning_rate": 0.00012302093023255814,
      "loss": 0.0566,
      "step": 3800
    },
    {
      "epoch": 74.07,
      "learning_rate": 0.00010767209302325581,
      "loss": 0.0506,
      "step": 4000
    },
    {
      "epoch": 74.07,
      "eval_loss": 1.2748003005981445,
      "eval_runtime": 12.9412,
      "eval_samples_per_second": 27.818,
      "eval_steps_per_second": 3.477,
      "eval_wer": 0.5507644061152489,
      "step": 4000
    },
    {
      "epoch": 77.78,
      "learning_rate": 9.232325581395348e-05,
      "loss": 0.0443,
      "step": 4200
    },
    {
      "epoch": 81.48,
      "learning_rate": 7.697441860465116e-05,
      "loss": 0.0397,
      "step": 4400
    },
    {
      "epoch": 83.33,
      "eval_loss": 1.1869040727615356,
      "eval_runtime": 12.9633,
      "eval_samples_per_second": 27.771,
      "eval_steps_per_second": 3.471,
      "eval_wer": 0.5393963151705213,
      "step": 4500
    },
    {
      "epoch": 85.19,
      "learning_rate": 6.162558139534884e-05,
      "loss": 0.0353,
      "step": 4600
    },
    {
      "epoch": 88.89,
      "learning_rate": 4.6276744186046515e-05,
      "loss": 0.0282,
      "step": 4800
    },
    {
      "epoch": 92.59,
      "learning_rate": 3.092790697674419e-05,
      "loss": 0.0276,
      "step": 5000
    },
    {
      "epoch": 92.59,
      "eval_loss": 1.2239253520965576,
      "eval_runtime": 12.8775,
      "eval_samples_per_second": 27.956,
      "eval_steps_per_second": 3.494,
      "eval_wer": 0.5390043120344963,
      "step": 5000
    },
    {
      "epoch": 96.3,
      "learning_rate": 1.5579069767441862e-05,
      "loss": 0.0247,
      "step": 5200
    },
    {
      "epoch": 100.0,
      "learning_rate": 2.302325581395349e-07,
      "loss": 0.0218,
      "step": 5400
    },
    {
      "epoch": 100.0,
      "step": 5400,
      "total_flos": 1.3559072525710848e+19,
      "train_loss": 0.7377449184876901,
      "train_runtime": 5722.2661,
      "train_samples_per_second": 14.959,
      "train_steps_per_second": 0.944
    }
  ],
  "max_steps": 5400,
  "num_train_epochs": 100,
  "total_flos": 1.3559072525710848e+19,
  "trial_name": null,
  "trial_params": null
}