trainer_state.json

{
  "best_metric": 1.0497090816497803,
  "best_model_checkpoint": "/home/labuser/Documents/phi-3/phi-3.5-new/checkpoint-63",
  "epoch": 8.952380952380953,
  "eval_steps": 500,
  "global_step": 94,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9523809523809523,
      "grad_norm": 2.605168342590332,
      "learning_rate": 0.0001,
      "loss": 1.515,
      "step": 10
    },
    {
      "epoch": 0.9523809523809523,
      "eval_loss": 1.310670256614685,
      "eval_runtime": 3.9865,
      "eval_samples_per_second": 4.264,
      "eval_steps_per_second": 2.258,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.2956364154815674,
      "learning_rate": 9.635919272833938e-05,
      "loss": 1.0908,
      "step": 21
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.1376752853393555,
      "eval_runtime": 3.9866,
      "eval_samples_per_second": 4.264,
      "eval_steps_per_second": 2.258,
      "step": 21
    },
    {
      "epoch": 2.9523809523809526,
      "grad_norm": 1.952481746673584,
      "learning_rate": 8.715724127386972e-05,
      "loss": 1.0257,
      "step": 31
    },
    {
      "epoch": 2.9523809523809526,
      "eval_loss": 1.0673834085464478,
      "eval_runtime": 4.0023,
      "eval_samples_per_second": 4.248,
      "eval_steps_per_second": 2.249,
      "step": 31
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.819559097290039,
      "learning_rate": 7.191855733945387e-05,
      "loss": 0.886,
      "step": 42
    },
    {
      "epoch": 4.0,
      "eval_loss": 1.050466537475586,
      "eval_runtime": 4.0072,
      "eval_samples_per_second": 4.242,
      "eval_steps_per_second": 2.246,
      "step": 42
    },
    {
      "epoch": 4.9523809523809526,
      "grad_norm": 2.7341201305389404,
      "learning_rate": 5.522642316338268e-05,
      "loss": 0.9453,
      "step": 52
    },
    {
      "epoch": 4.9523809523809526,
      "eval_loss": 1.0541181564331055,
      "eval_runtime": 4.0101,
      "eval_samples_per_second": 4.239,
      "eval_steps_per_second": 2.244,
      "step": 52
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.5818099975585938,
      "learning_rate": 3.6218132209150045e-05,
      "loss": 0.8409,
      "step": 63
    },
    {
      "epoch": 6.0,
      "eval_loss": 1.0497090816497803,
      "eval_runtime": 3.9956,
      "eval_samples_per_second": 4.255,
      "eval_steps_per_second": 2.252,
      "step": 63
    },
    {
      "epoch": 6.9523809523809526,
      "grad_norm": 2.11718487739563,
      "learning_rate": 2.061073738537635e-05,
      "loss": 0.9078,
      "step": 73
    },
    {
      "epoch": 6.9523809523809526,
      "eval_loss": 1.0589897632598877,
      "eval_runtime": 4.0003,
      "eval_samples_per_second": 4.25,
      "eval_steps_per_second": 2.25,
      "step": 73
    },
    {
      "epoch": 8.0,
      "grad_norm": 2.463954210281372,
      "learning_rate": 7.597595192178702e-06,
      "loss": 0.8154,
      "step": 84
    },
    {
      "epoch": 8.0,
      "eval_loss": 1.0593781471252441,
      "eval_runtime": 3.9931,
      "eval_samples_per_second": 4.257,
      "eval_steps_per_second": 2.254,
      "step": 84
    },
    {
      "epoch": 8.952380952380953,
      "grad_norm": 2.035900831222534,
      "learning_rate": 1.0926199633097157e-06,
      "loss": 0.8909,
      "step": 94
    },
    {
      "epoch": 8.952380952380953,
      "eval_loss": 1.057528018951416,
      "eval_runtime": 3.9065,
      "eval_samples_per_second": 4.352,
      "eval_steps_per_second": 2.304,
      "step": 94
    },
    {
      "epoch": 8.952380952380953,
      "step": 94,
      "total_flos": 8571389664559104.0,
      "train_loss": 0.9873519552514908,
      "train_runtime": 598.1578,
      "train_samples_per_second": 1.388,
      "train_steps_per_second": 0.167
    }
  ],
  "logging_steps": 500,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 3,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 3
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 8571389664559104.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}