trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 339,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15,
      "learning_rate": 4.262536873156342e-05,
      "loss": 0.0544,
      "step": 50
    },
    {
      "epoch": 0.29,
      "learning_rate": 3.5250737463126844e-05,
      "loss": 0.0046,
      "step": 100
    },
    {
      "epoch": 0.44,
      "learning_rate": 2.7876106194690264e-05,
      "loss": 0.0017,
      "step": 150
    },
    {
      "epoch": 0.59,
      "learning_rate": 2.0501474926253688e-05,
      "loss": 0.0008,
      "step": 200
    },
    {
      "epoch": 0.74,
      "learning_rate": 1.3126843657817109e-05,
      "loss": 0.0003,
      "step": 250
    },
    {
      "epoch": 0.88,
      "learning_rate": 5.752212389380531e-06,
      "loss": 0.0005,
      "step": 300
    },
    {
      "epoch": 1.0,
      "step": 339,
      "total_flos": 2741653307243520.0,
      "train_loss": 0.009268463998593412,
      "train_runtime": 186.2906,
      "train_samples_per_second": 116.34,
      "train_steps_per_second": 1.82
    }
  ],
  "max_steps": 339,
  "num_train_epochs": 1,
  "total_flos": 2741653307243520.0,
  "trial_name": null,
  "trial_params": null
}