trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "eval_steps": 500,
  "global_step": 240,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.546875,
      "learning_rate": 1.879473751206489e-05,
      "loss": 1.9176,
      "step": 48
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.3600443601608276,
      "eval_runtime": 6.1925,
      "eval_samples_per_second": 6.136,
      "eval_steps_per_second": 0.484,
      "step": 48
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.296875,
      "learning_rate": 1.4016954246529697e-05,
      "loss": 1.2058,
      "step": 96
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.1673846244812012,
      "eval_runtime": 6.1968,
      "eval_samples_per_second": 6.132,
      "eval_steps_per_second": 0.484,
      "step": 96
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.271484375,
      "learning_rate": 7.545145128592009e-06,
      "loss": 1.1141,
      "step": 144
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.1143308877944946,
      "eval_runtime": 6.1947,
      "eval_samples_per_second": 6.134,
      "eval_steps_per_second": 0.484,
      "step": 144
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.271484375,
      "learning_rate": 2.1085949060360654e-06,
      "loss": 1.0792,
      "step": 192
    },
    {
      "epoch": 4.0,
      "eval_loss": 1.0958216190338135,
      "eval_runtime": 6.1881,
      "eval_samples_per_second": 6.141,
      "eval_steps_per_second": 0.485,
      "step": 192
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.267578125,
      "learning_rate": 0.0,
      "loss": 1.0715,
      "step": 240
    },
    {
      "epoch": 5.0,
      "eval_loss": 1.0950061082839966,
      "eval_runtime": 6.1921,
      "eval_samples_per_second": 6.137,
      "eval_steps_per_second": 0.484,
      "step": 240
    },
    {
      "epoch": 5.0,
      "step": 240,
      "total_flos": 3.3325337630736384e+17,
      "train_loss": 1.27764892578125,
      "train_runtime": 2077.2222,
      "train_samples_per_second": 1.822,
      "train_steps_per_second": 0.116
    }
  ],
  "logging_steps": 500,
  "max_steps": 240,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 3.3325337630736384e+17,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}