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": 8.297603900150138,
      "learning_rate": 1e-05,
      "loss": 1.6795,
      "step": 1
    },
    {
      "epoch": 1.0,
      "grad_norm": 2.7714905692865806,
      "learning_rate": 1.7485107481711014e-05,
      "loss": 1.5018,
      "step": 5
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.746938943862915,
      "eval_runtime": 1.6252,
      "eval_samples_per_second": 4.922,
      "eval_steps_per_second": 0.615,
      "step": 5
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.355589361090063,
      "learning_rate": 6.453951129574644e-06,
      "loss": 1.1848,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.9052479267120361,
      "eval_runtime": 1.5781,
      "eval_samples_per_second": 5.069,
      "eval_steps_per_second": 0.634,
      "step": 10
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.8474736569347063,
      "learning_rate": 0.0,
      "loss": 0.9623,
      "step": 15
    },
    {
      "epoch": 3.0,
      "eval_loss": 1.9656569957733154,
      "eval_runtime": 1.5681,
      "eval_samples_per_second": 5.102,
      "eval_steps_per_second": 0.638,
      "step": 15
    },
    {
      "epoch": 3.0,
      "step": 15,
      "total_flos": 3088349921280.0,
      "train_loss": 1.22816797097524,
      "train_runtime": 294.6904,
      "train_samples_per_second": 3.176,
      "train_steps_per_second": 0.051
    }
  ],
  "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
}