trainer_state.json

{
  "best_global_step": 10,
  "best_metric": 0.6887128353118896,
  "best_model_checkpoint": "tiny_bert_rand_5_v2_rte/checkpoint-10",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 60,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.6591024994850159,
      "learning_rate": 4.9e-05,
      "loss": 0.7015,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5451263537906137,
      "eval_loss": 0.6887128353118896,
      "eval_runtime": 0.0978,
      "eval_samples_per_second": 2831.344,
      "eval_steps_per_second": 20.443,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.5431327223777771,
      "learning_rate": 4.8e-05,
      "loss": 0.6825,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.555956678700361,
      "eval_loss": 0.690912663936615,
      "eval_runtime": 0.0936,
      "eval_samples_per_second": 2958.692,
      "eval_steps_per_second": 21.362,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.7614234089851379,
      "learning_rate": 4.7e-05,
      "loss": 0.6577,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5595667870036101,
      "eval_loss": 0.7134617567062378,
      "eval_runtime": 0.101,
      "eval_samples_per_second": 2741.401,
      "eval_steps_per_second": 19.794,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.5347082614898682,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6153,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5342960288808665,
      "eval_loss": 0.7627834677696228,
      "eval_runtime": 0.1127,
      "eval_samples_per_second": 2457.952,
      "eval_steps_per_second": 17.747,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.2026875019073486,
      "learning_rate": 4.5e-05,
      "loss": 0.5533,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5342960288808665,
      "eval_loss": 0.8296673893928528,
      "eval_runtime": 0.0956,
      "eval_samples_per_second": 2896.417,
      "eval_steps_per_second": 20.913,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.170400619506836,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.4693,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4981949458483754,
      "eval_loss": 0.9874968528747559,
      "eval_runtime": 0.1016,
      "eval_samples_per_second": 2727.237,
      "eval_steps_per_second": 19.691,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 391780256993280.0,
      "train_loss": 0.6132701635360718,
      "train_runtime": 12.9315,
      "train_samples_per_second": 9627.626,
      "train_steps_per_second": 38.665
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "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": 391780256993280.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}