trainer_state.json

{
  "best_global_step": 264,
  "best_metric": 0.44867685437202454,
  "best_model_checkpoint": "tiny_bert_rand_5_v2_sst2/checkpoint-264",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 1584,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 9.064800262451172,
      "learning_rate": 4.9e-05,
      "loss": 0.427,
      "step": 264
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8165137614678899,
      "eval_loss": 0.44867685437202454,
      "eval_runtime": 0.2808,
      "eval_samples_per_second": 3105.024,
      "eval_steps_per_second": 14.243,
      "step": 264
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.39294958114624,
      "learning_rate": 4.8e-05,
      "loss": 0.2402,
      "step": 528
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7947247706422018,
      "eval_loss": 0.5559569001197815,
      "eval_runtime": 0.2662,
      "eval_samples_per_second": 3275.758,
      "eval_steps_per_second": 15.026,
      "step": 528
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.649034023284912,
      "learning_rate": 4.7e-05,
      "loss": 0.1856,
      "step": 792
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.805045871559633,
      "eval_loss": 0.5291202664375305,
      "eval_runtime": 0.2725,
      "eval_samples_per_second": 3199.63,
      "eval_steps_per_second": 14.677,
      "step": 792
    },
    {
      "epoch": 4.0,
      "grad_norm": 11.652801513671875,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.155,
      "step": 1056
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7958715596330275,
      "eval_loss": 0.5920218825340271,
      "eval_runtime": 0.2711,
      "eval_samples_per_second": 3217.0,
      "eval_steps_per_second": 14.757,
      "step": 1056
    },
    {
      "epoch": 5.0,
      "grad_norm": 9.776528358459473,
      "learning_rate": 4.5e-05,
      "loss": 0.1259,
      "step": 1320
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7935779816513762,
      "eval_loss": 0.637986958026886,
      "eval_runtime": 0.2896,
      "eval_samples_per_second": 3010.989,
      "eval_steps_per_second": 13.812,
      "step": 1320
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.283332824707031,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.1068,
      "step": 1584
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.801605504587156,
      "eval_loss": 0.7637522220611572,
      "eval_runtime": 0.275,
      "eval_samples_per_second": 3170.916,
      "eval_steps_per_second": 14.545,
      "step": 1584
    },
    {
      "epoch": 6.0,
      "step": 1584,
      "total_flos": 1.0596790573590528e+16,
      "train_loss": 0.20675967678879248,
      "train_runtime": 217.3543,
      "train_samples_per_second": 15492.91,
      "train_steps_per_second": 60.73
    }
  ],
  "logging_steps": 1,
  "max_steps": 13200,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 5
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 1.0596790573590528e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}