trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 30.0,
  "global_step": 240,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 12.5,
      "learning_rate": 0.0003,
      "loss": 71.8361,
      "step": 100
    },
    {
      "epoch": 12.5,
      "eval_cer": 0.10269482884195193,
      "eval_loss": 315.5455322265625,
      "eval_runtime": 68.0521,
      "eval_samples_per_second": 0.338,
      "eval_steps_per_second": 0.044,
      "eval_wer": 0.19291338582677164,
      "step": 100
    },
    {
      "epoch": 25.0,
      "learning_rate": 0.0003,
      "loss": 30.1145,
      "step": 200
    },
    {
      "epoch": 25.0,
      "eval_cer": 0.10196649672250546,
      "eval_loss": 392.1836242675781,
      "eval_runtime": 62.6919,
      "eval_samples_per_second": 0.367,
      "eval_steps_per_second": 0.048,
      "eval_wer": 0.1968503937007874,
      "step": 200
    },
    {
      "epoch": 30.0,
      "step": 240,
      "total_flos": 5.530990608198865e+17,
      "train_loss": 46.28341929117838,
      "train_runtime": 18940.2239,
      "train_samples_per_second": 0.098,
      "train_steps_per_second": 0.013
    }
  ],
  "max_steps": 240,
  "num_train_epochs": 30,
  "total_flos": 5.530990608198865e+17,
  "trial_name": null,
  "trial_params": null
}