trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 9.0,
  "eval_steps": 500,
  "global_step": 63,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.43,
      "learning_rate": 9.607381059352038e-07,
      "loss": 2.1176,
      "step": 10
    },
    {
      "epoch": 2.86,
      "learning_rate": 8.117449009293668e-07,
      "loss": 1.5678,
      "step": 20
    },
    {
      "epoch": 4.29,
      "learning_rate": 5.868240888334652e-07,
      "loss": 1.4173,
      "step": 30
    },
    {
      "epoch": 5.71,
      "learning_rate": 3.4075667487415785e-07,
      "loss": 1.342,
      "step": 40
    },
    {
      "epoch": 7.14,
      "learning_rate": 1.3347406408508694e-07,
      "loss": 1.3108,
      "step": 50
    },
    {
      "epoch": 8.57,
      "learning_rate": 1.5461356885461075e-08,
      "loss": 1.2987,
      "step": 60
    },
    {
      "epoch": 9.0,
      "step": 63,
      "total_flos": 8.187554675294208e+17,
      "train_loss": 1.4981852107577853,
      "train_runtime": 2910.9748,
      "train_samples_per_second": 9.69,
      "train_steps_per_second": 0.022
    }
  ],
  "logging_steps": 10,
  "max_steps": 63,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 21,
  "total_flos": 8.187554675294208e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}