trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.6921765208244324,
  "best_model_checkpoint": "tiny_bert_rand_5_v2_wnli/checkpoint-6",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.4666485786437988,
      "learning_rate": 4.9e-05,
      "loss": 0.7042,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7057108879089355,
      "eval_runtime": 0.0312,
      "eval_samples_per_second": 2274.169,
      "eval_steps_per_second": 32.031,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.5073485374450684,
      "learning_rate": 4.8e-05,
      "loss": 0.6983,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5352112676056338,
      "eval_loss": 0.6921765208244324,
      "eval_runtime": 0.0315,
      "eval_samples_per_second": 2255.344,
      "eval_steps_per_second": 31.765,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6022185683250427,
      "learning_rate": 4.7e-05,
      "loss": 0.6965,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.4225352112676056,
      "eval_loss": 0.699878990650177,
      "eval_runtime": 0.0322,
      "eval_samples_per_second": 2207.136,
      "eval_steps_per_second": 31.086,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6738465428352356,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6923,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4225352112676056,
      "eval_loss": 0.7091219425201416,
      "eval_runtime": 0.0321,
      "eval_samples_per_second": 2210.806,
      "eval_steps_per_second": 31.138,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.6243847012519836,
      "learning_rate": 4.5e-05,
      "loss": 0.6938,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.38028169014084506,
      "eval_loss": 0.7132483720779419,
      "eval_runtime": 0.03,
      "eval_samples_per_second": 2367.666,
      "eval_steps_per_second": 33.347,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.7892072796821594,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6933,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.7150639295578003,
      "eval_runtime": 0.03,
      "eval_samples_per_second": 2364.02,
      "eval_steps_per_second": 33.296,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.48970165848731995,
      "learning_rate": 4.3e-05,
      "loss": 0.6915,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7154490351676941,
      "eval_runtime": 0.0306,
      "eval_samples_per_second": 2323.769,
      "eval_steps_per_second": 32.729,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 116563804707840.0,
      "train_loss": 0.695692119144258,
      "train_runtime": 7.6587,
      "train_samples_per_second": 4145.634,
      "train_steps_per_second": 19.586
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "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": 116563804707840.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}