trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.7272727272727275,
  "eval_steps": 500,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.18,
      "grad_norm": 14.31835273388475,
      "learning_rate": 1e-05,
      "loss": 1.3372,
      "step": 1
    },
    {
      "epoch": 0.91,
      "grad_norm": 8.206390403360235,
      "learning_rate": 1.7485107481711014e-05,
      "loss": 1.3601,
      "step": 5
    },
    {
      "epoch": 0.91,
      "eval_loss": 1.155458688735962,
      "eval_runtime": 34.4213,
      "eval_samples_per_second": 12.086,
      "eval_steps_per_second": 0.755,
      "step": 5
    },
    {
      "epoch": 1.82,
      "grad_norm": 4.159649158422721,
      "learning_rate": 6.453951129574644e-06,
      "loss": 1.0001,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.8497009873390198,
      "eval_runtime": 34.1752,
      "eval_samples_per_second": 12.173,
      "eval_steps_per_second": 0.761,
      "step": 11
    },
    {
      "epoch": 2.73,
      "grad_norm": 3.4713231851764346,
      "learning_rate": 0.0,
      "loss": 0.7357,
      "step": 15
    },
    {
      "epoch": 2.73,
      "eval_loss": 0.8050652146339417,
      "eval_runtime": 34.1928,
      "eval_samples_per_second": 12.166,
      "eval_steps_per_second": 0.76,
      "step": 15
    },
    {
      "epoch": 2.73,
      "step": 15,
      "total_flos": 3088349921280.0,
      "train_loss": 1.0304591496785482,
      "train_runtime": 361.8724,
      "train_samples_per_second": 2.852,
      "train_steps_per_second": 0.041
    }
  ],
  "logging_steps": 5,
  "max_steps": 15,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3088349921280.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}