trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 8472,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.18,
      "learning_rate": 4.7049102927289904e-05,
      "loss": 2.4017,
      "step": 500
    },
    {
      "epoch": 0.35,
      "learning_rate": 4.409820585457979e-05,
      "loss": 1.5917,
      "step": 1000
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.114730878186969e-05,
      "loss": 1.385,
      "step": 1500
    },
    {
      "epoch": 0.71,
      "learning_rate": 3.819641170915959e-05,
      "loss": 1.2751,
      "step": 2000
    },
    {
      "epoch": 0.89,
      "learning_rate": 3.524551463644948e-05,
      "loss": 1.1939,
      "step": 2500
    },
    {
      "epoch": 1.06,
      "learning_rate": 3.229461756373938e-05,
      "loss": 1.0865,
      "step": 3000
    },
    {
      "epoch": 1.24,
      "learning_rate": 2.9343720491029275e-05,
      "loss": 0.9701,
      "step": 3500
    },
    {
      "epoch": 1.42,
      "learning_rate": 2.639282341831917e-05,
      "loss": 0.9544,
      "step": 4000
    },
    {
      "epoch": 1.59,
      "learning_rate": 2.3441926345609068e-05,
      "loss": 0.9205,
      "step": 4500
    },
    {
      "epoch": 1.77,
      "learning_rate": 2.0491029272898962e-05,
      "loss": 0.9088,
      "step": 5000
    },
    {
      "epoch": 1.95,
      "learning_rate": 1.7540132200188857e-05,
      "loss": 0.8795,
      "step": 5500
    },
    {
      "epoch": 2.12,
      "learning_rate": 1.4589235127478753e-05,
      "loss": 0.7818,
      "step": 6000
    },
    {
      "epoch": 2.3,
      "learning_rate": 1.1638338054768651e-05,
      "loss": 0.733,
      "step": 6500
    },
    {
      "epoch": 2.48,
      "learning_rate": 8.687440982058546e-06,
      "loss": 0.728,
      "step": 7000
    },
    {
      "epoch": 2.66,
      "learning_rate": 5.736543909348442e-06,
      "loss": 0.7091,
      "step": 7500
    },
    {
      "epoch": 2.83,
      "learning_rate": 2.785646836638338e-06,
      "loss": 0.7155,
      "step": 8000
    },
    {
      "epoch": 3.0,
      "step": 8472,
      "total_flos": 1.362358015279104e+16,
      "train_loss": 1.0567845856952036,
      "train_runtime": 2844.4841,
      "train_samples_per_second": 29.783,
      "train_steps_per_second": 2.978
    }
  ],
  "max_steps": 8472,
  "num_train_epochs": 3,
  "total_flos": 1.362358015279104e+16,
  "trial_name": null,
  "trial_params": null
}