checkpoint-30/trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 10.402471542358398,
  "best_model_checkpoint": "models/phi35-mini-nlu/checkpoint-30",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 30,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.8480287790298462,
      "learning_rate": 3.6e-05,
      "loss": 11.401779174804688,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_loss": 11.322081565856934,
      "eval_runtime": 3.1229,
      "eval_samples_per_second": 5.444,
      "eval_steps_per_second": 1.601,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": NaN,
      "learning_rate": 7.6e-05,
      "loss": 11.099254608154297,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_loss": 10.799253463745117,
      "eval_runtime": 3.115,
      "eval_samples_per_second": 5.457,
      "eval_steps_per_second": 1.605,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": NaN,
      "learning_rate": 0.000116,
      "loss": 10.584548187255859,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_loss": 10.402471542358398,
      "eval_runtime": 3.1141,
      "eval_samples_per_second": 5.459,
      "eval_steps_per_second": 1.606,
      "step": 30
    }
  ],
  "logging_steps": 10,
  "max_steps": 30,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.0506901244608512e+16,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}