trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.957983193277311,
  "eval_steps": 500,
  "global_step": 132,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.22408963585434175,
      "grad_norm": 31.895059346549615,
      "learning_rate": 5e-06,
      "loss": 1.0566,
      "step": 10
    },
    {
      "epoch": 0.4481792717086835,
      "grad_norm": 1.7397235388731687,
      "learning_rate": 5e-06,
      "loss": 0.9929,
      "step": 20
    },
    {
      "epoch": 0.6722689075630253,
      "grad_norm": 24.73289602340893,
      "learning_rate": 5e-06,
      "loss": 0.9455,
      "step": 30
    },
    {
      "epoch": 0.896358543417367,
      "grad_norm": 15.103388080469578,
      "learning_rate": 5e-06,
      "loss": 0.9223,
      "step": 40
    },
    {
      "epoch": 0.9859943977591037,
      "eval_loss": 0.9011093974113464,
      "eval_runtime": 30.3519,
      "eval_samples_per_second": 39.569,
      "eval_steps_per_second": 0.626,
      "step": 44
    },
    {
      "epoch": 1.1204481792717087,
      "grad_norm": 1.3240129819926922,
      "learning_rate": 5e-06,
      "loss": 0.9194,
      "step": 50
    },
    {
      "epoch": 1.3445378151260505,
      "grad_norm": 1.0063155415585696,
      "learning_rate": 5e-06,
      "loss": 0.8587,
      "step": 60
    },
    {
      "epoch": 1.5686274509803921,
      "grad_norm": 1.087235116946335,
      "learning_rate": 5e-06,
      "loss": 0.8482,
      "step": 70
    },
    {
      "epoch": 1.7927170868347337,
      "grad_norm": 0.763889264805126,
      "learning_rate": 5e-06,
      "loss": 0.8393,
      "step": 80
    },
    {
      "epoch": 1.9943977591036415,
      "eval_loss": 0.8523734211921692,
      "eval_runtime": 30.796,
      "eval_samples_per_second": 38.999,
      "eval_steps_per_second": 0.617,
      "step": 89
    },
    {
      "epoch": 2.0168067226890756,
      "grad_norm": 1.0022099725803806,
      "learning_rate": 5e-06,
      "loss": 0.8585,
      "step": 90
    },
    {
      "epoch": 2.2408963585434174,
      "grad_norm": 0.8920948411419858,
      "learning_rate": 5e-06,
      "loss": 0.7773,
      "step": 100
    },
    {
      "epoch": 2.4649859943977592,
      "grad_norm": 0.7868621725302428,
      "learning_rate": 5e-06,
      "loss": 0.7786,
      "step": 110
    },
    {
      "epoch": 2.689075630252101,
      "grad_norm": 0.7973258054338478,
      "learning_rate": 5e-06,
      "loss": 0.7664,
      "step": 120
    },
    {
      "epoch": 2.9131652661064424,
      "grad_norm": 0.8414953772946839,
      "learning_rate": 5e-06,
      "loss": 0.7699,
      "step": 130
    },
    {
      "epoch": 2.957983193277311,
      "eval_loss": 0.8426499366760254,
      "eval_runtime": 29.0845,
      "eval_samples_per_second": 41.293,
      "eval_steps_per_second": 0.653,
      "step": 132
    },
    {
      "epoch": 2.957983193277311,
      "step": 132,
      "total_flos": 220895536742400.0,
      "train_loss": 0.8702321576349663,
      "train_runtime": 4622.8871,
      "train_samples_per_second": 14.805,
      "train_steps_per_second": 0.029
    }
  ],
  "logging_steps": 10,
  "max_steps": 132,
  "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": 220895536742400.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}