trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6398081183433533,
  "best_model_checkpoint": "tiny_bert_rand_20_v1_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.045642614364624,
      "learning_rate": 4.9e-05,
      "loss": 0.6624,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6293245469522241,
      "eval_loss": 0.6432347893714905,
      "eval_runtime": 1.6847,
      "eval_samples_per_second": 3242.7,
      "eval_steps_per_second": 13.059,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.2814295291900635,
      "learning_rate": 4.8e-05,
      "loss": 0.6204,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6339007871133077,
      "eval_loss": 0.6398081183433533,
      "eval_runtime": 1.6272,
      "eval_samples_per_second": 3357.237,
      "eval_steps_per_second": 13.52,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.44016170501709,
      "learning_rate": 4.7e-05,
      "loss": 0.5434,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6359143327841845,
      "eval_loss": 0.663581907749176,
      "eval_runtime": 1.5971,
      "eval_samples_per_second": 3420.507,
      "eval_steps_per_second": 13.775,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.1715989112854,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4486,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6377448288486179,
      "eval_loss": 0.7344562411308289,
      "eval_runtime": 1.6198,
      "eval_samples_per_second": 3372.664,
      "eval_steps_per_second": 13.582,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 7.201107025146484,
      "learning_rate": 4.5e-05,
      "loss": 0.3598,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6315211422295443,
      "eval_loss": 0.8187642693519592,
      "eval_runtime": 1.6163,
      "eval_samples_per_second": 3379.985,
      "eval_steps_per_second": 13.612,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.800537109375,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2863,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.633534687900421,
      "eval_loss": 0.9331047534942627,
      "eval_runtime": 1.6228,
      "eval_samples_per_second": 3366.391,
      "eval_steps_per_second": 13.557,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 12.566457748413086,
      "learning_rate": 4.3e-05,
      "loss": 0.228,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.629873695771554,
      "eval_loss": 1.1205413341522217,
      "eval_runtime": 1.6178,
      "eval_samples_per_second": 3376.813,
      "eval_steps_per_second": 13.599,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 1.922715369529651e+16,
      "train_loss": 0.4498554375944237,
      "train_runtime": 403.8539,
      "train_samples_per_second": 12967.931,
      "train_steps_per_second": 50.761
    }
  ],
  "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": 1.922715369529651e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}