trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 736,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.54,
      "grad_norm": 1.9390450716018677,
      "learning_rate": 0.00017282608695652174,
      "loss": 1.766,
      "step": 100
    },
    {
      "epoch": 1.09,
      "grad_norm": 1.6820060014724731,
      "learning_rate": 0.0001456521739130435,
      "loss": 1.4593,
      "step": 200
    },
    {
      "epoch": 1.63,
      "grad_norm": 1.9340823888778687,
      "learning_rate": 0.00011847826086956522,
      "loss": 1.3828,
      "step": 300
    },
    {
      "epoch": 2.17,
      "grad_norm": 2.359797477722168,
      "learning_rate": 9.130434782608696e-05,
      "loss": 1.348,
      "step": 400
    },
    {
      "epoch": 2.72,
      "grad_norm": 2.2974631786346436,
      "learning_rate": 6.413043478260869e-05,
      "loss": 1.288,
      "step": 500
    },
    {
      "epoch": 3.26,
      "grad_norm": 2.135037422180176,
      "learning_rate": 3.695652173913043e-05,
      "loss": 1.2496,
      "step": 600
    },
    {
      "epoch": 3.8,
      "grad_norm": 2.4940547943115234,
      "learning_rate": 9.782608695652175e-06,
      "loss": 1.2264,
      "step": 700
    }
  ],
  "logging_steps": 100,
  "max_steps": 736,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 1.846229215266048e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}