trainer_state.json

{
  "best_metric": 2.911093627065303,
  "best_model_checkpoint": "./3data-finetune/checkpoint-200",
  "epoch": 2.2222222222222223,
  "eval_steps": 25,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.28,
      "learning_rate": 5e-06,
      "loss": 0.2226,
      "step": 25
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.08195594698190689,
      "eval_runtime": 247.5945,
      "eval_samples_per_second": 1.026,
      "eval_steps_per_second": 0.513,
      "eval_wer": 6.058221872541306,
      "step": 25
    },
    {
      "epoch": 0.56,
      "learning_rate": 9.800000000000001e-06,
      "loss": 0.079,
      "step": 50
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.07928125560283661,
      "eval_runtime": 245.6975,
      "eval_samples_per_second": 1.034,
      "eval_steps_per_second": 0.517,
      "eval_wer": 5.664830841856806,
      "step": 50
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.400000000000001e-06,
      "loss": 0.0773,
      "step": 75
    },
    {
      "epoch": 0.83,
      "eval_loss": 0.06937055289745331,
      "eval_runtime": 247.4688,
      "eval_samples_per_second": 1.026,
      "eval_steps_per_second": 0.513,
      "eval_wer": 4.248623131392605,
      "step": 75
    },
    {
      "epoch": 1.11,
      "learning_rate": 6.800000000000001e-06,
      "loss": 0.0624,
      "step": 100
    },
    {
      "epoch": 1.11,
      "eval_loss": 0.059788815677165985,
      "eval_runtime": 245.9389,
      "eval_samples_per_second": 1.033,
      "eval_steps_per_second": 0.516,
      "eval_wer": 3.6191974822974036,
      "step": 100
    },
    {
      "epoch": 1.39,
      "learning_rate": 5.133333333333334e-06,
      "loss": 0.0426,
      "step": 125
    },
    {
      "epoch": 1.39,
      "eval_loss": 0.055591508746147156,
      "eval_runtime": 246.0647,
      "eval_samples_per_second": 1.032,
      "eval_steps_per_second": 0.516,
      "eval_wer": 2.832415420928403,
      "step": 125
    },
    {
      "epoch": 1.67,
      "learning_rate": 3.4666666666666672e-06,
      "loss": 0.0575,
      "step": 150
    },
    {
      "epoch": 1.67,
      "eval_loss": 0.05484224110841751,
      "eval_runtime": 248.469,
      "eval_samples_per_second": 1.022,
      "eval_steps_per_second": 0.511,
      "eval_wer": 3.6978756884343036,
      "step": 150
    },
    {
      "epoch": 1.94,
      "learning_rate": 1.8000000000000001e-06,
      "loss": 0.0312,
      "step": 175
    },
    {
      "epoch": 1.94,
      "eval_loss": 0.05231180042028427,
      "eval_runtime": 246.3108,
      "eval_samples_per_second": 1.031,
      "eval_steps_per_second": 0.516,
      "eval_wer": 2.911093627065303,
      "step": 175
    },
    {
      "epoch": 2.22,
      "learning_rate": 1.3333333333333336e-07,
      "loss": 0.0236,
      "step": 200
    },
    {
      "epoch": 2.22,
      "eval_loss": 0.052269306033849716,
      "eval_runtime": 248.7723,
      "eval_samples_per_second": 1.021,
      "eval_steps_per_second": 0.511,
      "eval_wer": 2.911093627065303,
      "step": 200
    }
  ],
  "logging_steps": 25,
  "max_steps": 200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 50,
  "total_flos": 3.265935704064e+18,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}