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": 8.112652061034474,
      "learning_rate": 1e-05,
      "loss": 1.4798,
      "step": 1
    },
    {
      "epoch": 0.91,
      "grad_norm": 2.848398983756101,
      "learning_rate": 1.7485107481711014e-05,
      "loss": 1.2659,
      "step": 5
    },
    {
      "epoch": 0.91,
      "eval_loss": 1.2053821086883545,
      "eval_runtime": 1.8264,
      "eval_samples_per_second": 3.833,
      "eval_steps_per_second": 0.548,
      "step": 5
    },
    {
      "epoch": 1.82,
      "grad_norm": 2.6874796106077996,
      "learning_rate": 6.453951129574644e-06,
      "loss": 1.0094,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.1690208911895752,
      "eval_runtime": 1.8589,
      "eval_samples_per_second": 3.766,
      "eval_steps_per_second": 0.538,
      "step": 11
    },
    {
      "epoch": 2.73,
      "grad_norm": 1.7499037267304558,
      "learning_rate": 0.0,
      "loss": 0.8249,
      "step": 15
    },
    {
      "epoch": 2.73,
      "eval_loss": 1.1581600904464722,
      "eval_runtime": 1.8666,
      "eval_samples_per_second": 3.75,
      "eval_steps_per_second": 0.536,
      "step": 15
    },
    {
      "epoch": 2.73,
      "step": 15,
      "total_flos": 3088349921280.0,
      "train_loss": 1.0476287603378296,
      "train_runtime": 318.5273,
      "train_samples_per_second": 3.306,
      "train_steps_per_second": 0.047
    }
  ],
  "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
}