trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 31.25,
  "eval_steps": 500,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.56,
      "grad_norm": 1.203125,
      "learning_rate": 3e-05,
      "loss": 1.3475,
      "step": 25
    },
    {
      "epoch": 3.12,
      "grad_norm": 1.6328125,
      "learning_rate": 3e-05,
      "loss": 0.944,
      "step": 50
    },
    {
      "epoch": 4.69,
      "grad_norm": 2.390625,
      "learning_rate": 3e-05,
      "loss": 0.6433,
      "step": 75
    },
    {
      "epoch": 6.25,
      "grad_norm": 2.953125,
      "learning_rate": 3e-05,
      "loss": 0.3333,
      "step": 100
    },
    {
      "epoch": 7.81,
      "grad_norm": 2.453125,
      "learning_rate": 3e-05,
      "loss": 0.1399,
      "step": 125
    },
    {
      "epoch": 9.38,
      "grad_norm": 2.375,
      "learning_rate": 3e-05,
      "loss": 0.0672,
      "step": 150
    },
    {
      "epoch": 10.94,
      "grad_norm": 2.453125,
      "learning_rate": 3e-05,
      "loss": 0.0431,
      "step": 175
    },
    {
      "epoch": 12.5,
      "grad_norm": 2.546875,
      "learning_rate": 3e-05,
      "loss": 0.0302,
      "step": 200
    },
    {
      "epoch": 14.06,
      "grad_norm": 1.578125,
      "learning_rate": 3e-05,
      "loss": 0.0276,
      "step": 225
    },
    {
      "epoch": 15.62,
      "grad_norm": 1.75,
      "learning_rate": 3e-05,
      "loss": 0.0236,
      "step": 250
    },
    {
      "epoch": 17.19,
      "grad_norm": 1.359375,
      "learning_rate": 3e-05,
      "loss": 0.0223,
      "step": 275
    },
    {
      "epoch": 18.75,
      "grad_norm": 1.328125,
      "learning_rate": 3e-05,
      "loss": 0.0203,
      "step": 300
    },
    {
      "epoch": 20.31,
      "grad_norm": 0.87109375,
      "learning_rate": 3e-05,
      "loss": 0.0191,
      "step": 325
    },
    {
      "epoch": 21.88,
      "grad_norm": 1.3828125,
      "learning_rate": 3e-05,
      "loss": 0.0186,
      "step": 350
    },
    {
      "epoch": 23.44,
      "grad_norm": 1.5078125,
      "learning_rate": 3e-05,
      "loss": 0.0164,
      "step": 375
    },
    {
      "epoch": 25.0,
      "grad_norm": 1.9765625,
      "learning_rate": 3e-05,
      "loss": 0.0172,
      "step": 400
    },
    {
      "epoch": 26.56,
      "grad_norm": 1.234375,
      "learning_rate": 3e-05,
      "loss": 0.0148,
      "step": 425
    },
    {
      "epoch": 28.12,
      "grad_norm": 0.482421875,
      "learning_rate": 3e-05,
      "loss": 0.0154,
      "step": 450
    },
    {
      "epoch": 29.69,
      "grad_norm": 1.0625,
      "learning_rate": 3e-05,
      "loss": 0.0143,
      "step": 475
    },
    {
      "epoch": 31.25,
      "grad_norm": 0.8125,
      "learning_rate": 3e-05,
      "loss": 0.0144,
      "step": 500
    }
  ],
  "logging_steps": 25,
  "max_steps": 500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 32,
  "save_steps": 50,
  "total_flos": 1.5640413609804595e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}