trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 20,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5,
      "grad_norm": 17.86797523498535,
      "learning_rate": 1e-05,
      "loss": 1.1144,
      "step": 1
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.2928184270858765,
      "eval_runtime": 5.1117,
      "eval_samples_per_second": 9.39,
      "eval_steps_per_second": 0.391,
      "step": 2
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.8672829270362854,
      "eval_runtime": 5.1217,
      "eval_samples_per_second": 9.372,
      "eval_steps_per_second": 0.39,
      "step": 4
    },
    {
      "epoch": 2.5,
      "grad_norm": 25.679489135742188,
      "learning_rate": 8.94570254698197e-06,
      "loss": 1.0902,
      "step": 5
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.6943734288215637,
      "eval_runtime": 5.1377,
      "eval_samples_per_second": 9.343,
      "eval_steps_per_second": 0.389,
      "step": 6
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.4531029760837555,
      "eval_runtime": 5.1394,
      "eval_samples_per_second": 9.34,
      "eval_steps_per_second": 0.389,
      "step": 8
    },
    {
      "epoch": 5.0,
      "grad_norm": 21.12933349609375,
      "learning_rate": 5.412896727361663e-06,
      "loss": 0.5758,
      "step": 10
    },
    {
      "epoch": 5.0,
      "eval_loss": 0.31279101967811584,
      "eval_runtime": 5.1429,
      "eval_samples_per_second": 9.333,
      "eval_steps_per_second": 0.389,
      "step": 10
    },
    {
      "epoch": 6.0,
      "eval_loss": 0.18658435344696045,
      "eval_runtime": 5.1385,
      "eval_samples_per_second": 9.341,
      "eval_steps_per_second": 0.389,
      "step": 12
    },
    {
      "epoch": 7.0,
      "eval_loss": 0.1056341603398323,
      "eval_runtime": 5.1413,
      "eval_samples_per_second": 9.336,
      "eval_steps_per_second": 0.389,
      "step": 14
    },
    {
      "epoch": 7.5,
      "grad_norm": 2.3144567012786865,
      "learning_rate": 1.6135921418712959e-06,
      "loss": 0.1998,
      "step": 15
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.0681278184056282,
      "eval_runtime": 5.1435,
      "eval_samples_per_second": 9.332,
      "eval_steps_per_second": 0.389,
      "step": 16
    },
    {
      "epoch": 9.0,
      "eval_loss": 0.05328406020998955,
      "eval_runtime": 5.1445,
      "eval_samples_per_second": 9.33,
      "eval_steps_per_second": 0.389,
      "step": 18
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.7936142086982727,
      "learning_rate": 0.0,
      "loss": 0.0645,
      "step": 20
    },
    {
      "epoch": 10.0,
      "eval_loss": 0.05004188045859337,
      "eval_runtime": 5.0849,
      "eval_samples_per_second": 9.44,
      "eval_steps_per_second": 0.393,
      "step": 20
    },
    {
      "epoch": 10.0,
      "step": 20,
      "total_flos": 1.1184101603357491e+17,
      "train_loss": 0.4837849587202072,
      "train_runtime": 336.0147,
      "train_samples_per_second": 1.429,
      "train_steps_per_second": 0.06
    }
  ],
  "logging_steps": 5,
  "max_steps": 20,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.1184101603357491e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}