trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "grad_norm": 7.938874100772174,
      "learning_rate": 1e-05,
      "loss": 1.4592,
      "step": 1
    },
    {
      "epoch": 1.0,
      "grad_norm": 2.9088166904840036,
      "learning_rate": 1.7485107481711014e-05,
      "loss": 1.2724,
      "step": 5
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.2259724140167236,
      "eval_runtime": 5.3259,
      "eval_samples_per_second": 8.449,
      "eval_steps_per_second": 0.563,
      "step": 5
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.8826257662011012,
      "learning_rate": 6.453951129574644e-06,
      "loss": 0.9804,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.1944764852523804,
      "eval_runtime": 5.4002,
      "eval_samples_per_second": 8.333,
      "eval_steps_per_second": 0.556,
      "step": 10
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.735060327966447,
      "learning_rate": 0.0,
      "loss": 0.781,
      "step": 15
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.1811635494232178,
      "eval_runtime": 5.3908,
      "eval_samples_per_second": 8.348,
      "eval_steps_per_second": 0.557,
      "step": 15
    },
    {
      "epoch": 3.0,
      "step": 15,
      "total_flos": 3088349921280.0,
      "train_loss": 1.0237295468648275,
      "train_runtime": 322.7929,
      "train_samples_per_second": 2.918,
      "train_steps_per_second": 0.046
    }
  ],
  "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
}