trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9985772550507446,
  "eval_steps": 500,
  "global_step": 658,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15,
      "grad_norm": 1.3818485736846924,
      "learning_rate": 1.984705192213004e-05,
      "loss": 1.5534,
      "step": 100
    },
    {
      "epoch": 0.3,
      "grad_norm": 1.1903014183044434,
      "learning_rate": 1.761094828138534e-05,
      "loss": 0.8705,
      "step": 200
    },
    {
      "epoch": 0.46,
      "grad_norm": 1.446919322013855,
      "learning_rate": 1.3281315178199537e-05,
      "loss": 0.8352,
      "step": 300
    },
    {
      "epoch": 0.61,
      "grad_norm": 1.4014055728912354,
      "learning_rate": 8.04909677983872e-06,
      "loss": 0.8368,
      "step": 400
    },
    {
      "epoch": 0.76,
      "grad_norm": 1.499550461769104,
      "learning_rate": 3.3535097208036584e-06,
      "loss": 0.8154,
      "step": 500
    },
    {
      "epoch": 0.91,
      "grad_norm": 1.5235815048217773,
      "learning_rate": 4.861604459298697e-07,
      "loss": 0.8083,
      "step": 600
    },
    {
      "epoch": 1.0,
      "step": 658,
      "total_flos": 4.878623526912e+16,
      "train_loss": 0.9400251382752393,
      "train_runtime": 7654.5047,
      "train_samples_per_second": 1.377,
      "train_steps_per_second": 0.086
    }
  ],
  "logging_steps": 100,
  "max_steps": 658,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 4.878623526912e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}