trainer_state.json

{
  "best_global_step": 410,
  "best_metric": 0.64823979139328,
  "best_model_checkpoint": "tiny_bert_rand_10_v1_qnli/checkpoint-410",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 2460,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.5401153564453125,
      "learning_rate": 4.9e-05,
      "loss": 0.6652,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6110195863078894,
      "eval_loss": 0.64823979139328,
      "eval_runtime": 1.6261,
      "eval_samples_per_second": 3359.661,
      "eval_steps_per_second": 13.53,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6169238090515137,
      "learning_rate": 4.8e-05,
      "loss": 0.6354,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6181585209591799,
      "eval_loss": 0.6503065824508667,
      "eval_runtime": 1.5955,
      "eval_samples_per_second": 3424.057,
      "eval_steps_per_second": 13.789,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.669760227203369,
      "learning_rate": 4.7e-05,
      "loss": 0.5927,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6199890170236134,
      "eval_loss": 0.6662376523017883,
      "eval_runtime": 1.6111,
      "eval_samples_per_second": 3390.858,
      "eval_steps_per_second": 13.655,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.882028579711914,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.5314,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6135822807980963,
      "eval_loss": 0.7151873111724854,
      "eval_runtime": 1.6049,
      "eval_samples_per_second": 3403.879,
      "eval_steps_per_second": 13.708,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 4.443440914154053,
      "learning_rate": 4.5e-05,
      "loss": 0.4613,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6055280981145891,
      "eval_loss": 0.7752958536148071,
      "eval_runtime": 1.6021,
      "eval_samples_per_second": 3409.898,
      "eval_steps_per_second": 13.732,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.031154155731201,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.3887,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.6093721398498994,
      "eval_loss": 0.9132418036460876,
      "eval_runtime": 1.6082,
      "eval_samples_per_second": 3397.02,
      "eval_steps_per_second": 13.68,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "step": 2460,
      "total_flos": 1.6480417453111296e+16,
      "train_loss": 0.5457534603956269,
      "train_runtime": 346.5609,
      "train_samples_per_second": 15111.775,
      "train_steps_per_second": 59.153
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 1.6480417453111296e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}