last-checkpoint/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 20.0,
  "eval_steps": 200,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "eval_loss": 6.9372711181640625,
      "eval_runtime": 0.455,
      "eval_samples_per_second": 3292.619,
      "eval_steps_per_second": 26.376,
      "step": 1
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.84375,
      "learning_rate": 0.0002,
      "loss": 6.9316,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.58203125,
      "learning_rate": 0.00019863613034027224,
      "loss": 6.882,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.359375,
      "learning_rate": 0.00019458172417006347,
      "loss": 6.8025,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.34375,
      "learning_rate": 0.0001879473751206489,
      "loss": 6.7272,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.33984375,
      "learning_rate": 0.00017891405093963938,
      "loss": 6.6579,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.333984375,
      "learning_rate": 0.00016772815716257412,
      "loss": 6.5946,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.328125,
      "learning_rate": 0.00015469481581224272,
      "loss": 6.5398,
      "step": 70
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.326171875,
      "learning_rate": 0.00014016954246529696,
      "loss": 6.4928,
      "step": 80
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.322265625,
      "learning_rate": 0.00012454854871407994,
      "loss": 6.4527,
      "step": 90
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.3203125,
      "learning_rate": 0.00010825793454723325,
      "loss": 6.4192,
      "step": 100
    },
    {
      "epoch": 11.0,
      "grad_norm": 0.318359375,
      "learning_rate": 9.174206545276677e-05,
      "loss": 6.3933,
      "step": 110
    },
    {
      "epoch": 12.0,
      "grad_norm": 0.3125,
      "learning_rate": 7.54514512859201e-05,
      "loss": 6.3733,
      "step": 120
    },
    {
      "epoch": 13.0,
      "grad_norm": 0.30859375,
      "learning_rate": 5.983045753470308e-05,
      "loss": 6.3581,
      "step": 130
    },
    {
      "epoch": 14.0,
      "grad_norm": 0.310546875,
      "learning_rate": 4.530518418775733e-05,
      "loss": 6.3482,
      "step": 140
    },
    {
      "epoch": 15.0,
      "grad_norm": 0.314453125,
      "learning_rate": 3.227184283742591e-05,
      "loss": 6.3422,
      "step": 150
    },
    {
      "epoch": 16.0,
      "grad_norm": 0.310546875,
      "learning_rate": 2.1085949060360654e-05,
      "loss": 6.3393,
      "step": 160
    },
    {
      "epoch": 17.0,
      "grad_norm": 0.3125,
      "learning_rate": 1.2052624879351104e-05,
      "loss": 6.3382,
      "step": 170
    },
    {
      "epoch": 18.0,
      "grad_norm": 0.30859375,
      "learning_rate": 5.418275829936537e-06,
      "loss": 6.3376,
      "step": 180
    },
    {
      "epoch": 19.0,
      "grad_norm": 0.310546875,
      "learning_rate": 1.3638696597277679e-06,
      "loss": 6.3375,
      "step": 190
    },
    {
      "epoch": 20.0,
      "grad_norm": 0.30859375,
      "learning_rate": 0.0,
      "loss": 6.3375,
      "step": 200
    },
    {
      "epoch": 20.0,
      "eval_loss": 6.335957050323486,
      "eval_runtime": 0.4212,
      "eval_samples_per_second": 3556.438,
      "eval_steps_per_second": 28.489,
      "step": 200
    }
  ],
  "logging_steps": 10,
  "max_steps": 200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 20,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 14891561779200.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}