trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.02139495079161318,
  "global_step": 150,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 4.0000000000000003e-07,
      "loss": 1.3206,
      "step": 25
    },
    {
      "epoch": 0.01,
      "learning_rate": 8.8e-07,
      "loss": 1.247,
      "step": 50
    },
    {
      "epoch": 0.01,
      "learning_rate": 1.3800000000000001e-06,
      "loss": 1.1224,
      "step": 75
    },
    {
      "epoch": 0.01,
      "learning_rate": 1.8800000000000002e-06,
      "loss": 0.9567,
      "step": 100
    },
    {
      "epoch": 0.02,
      "learning_rate": 2.38e-06,
      "loss": 0.964,
      "step": 125
    },
    {
      "epoch": 0.02,
      "learning_rate": 2.88e-06,
      "loss": 0.7427,
      "step": 150
    },
    {
      "epoch": 0.02,
      "step": 150,
      "total_flos": 9728999424000000.0,
      "train_loss": 1.0588922500610352,
      "train_runtime": 31.6687,
      "train_samples_per_second": 4.737,
      "train_steps_per_second": 4.737
    },
    {
      "epoch": 0.02,
      "eval_cer": 16.883464889935755,
      "eval_loss": 0.8340538144111633,
      "eval_runtime": 2688.9813,
      "eval_samples_per_second": 0.727,
      "eval_steps_per_second": 0.363,
      "eval_wer": 41.865416504595984,
      "step": 150
    }
  ],
  "max_steps": 150,
  "num_train_epochs": 1,
  "total_flos": 9728999424000000.0,
  "trial_name": null,
  "trial_params": null
}