trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.7452685236930847,
  "best_model_checkpoint": "tiny_bert_rand_20_v1_wnli/checkpoint-3",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.0748443603515625,
      "learning_rate": 4.9e-05,
      "loss": 0.7076,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.30985915492957744,
      "eval_loss": 0.7452685236930847,
      "eval_runtime": 0.0299,
      "eval_samples_per_second": 2377.496,
      "eval_steps_per_second": 33.486,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8317832946777344,
      "learning_rate": 4.8e-05,
      "loss": 0.6916,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.7565469145774841,
      "eval_runtime": 0.0312,
      "eval_samples_per_second": 2279.164,
      "eval_steps_per_second": 32.101,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.802528440952301,
      "learning_rate": 4.7e-05,
      "loss": 0.6861,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.22535211267605634,
      "eval_loss": 0.778829038143158,
      "eval_runtime": 0.0333,
      "eval_samples_per_second": 2130.474,
      "eval_steps_per_second": 30.007,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.9218354821205139,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6797,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.14084507042253522,
      "eval_loss": 0.8008360266685486,
      "eval_runtime": 0.0341,
      "eval_samples_per_second": 2085.053,
      "eval_steps_per_second": 29.367,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.9836104512214661,
      "learning_rate": 4.5e-05,
      "loss": 0.6846,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.15492957746478872,
      "eval_loss": 0.8251541256904602,
      "eval_runtime": 0.0305,
      "eval_samples_per_second": 2329.239,
      "eval_steps_per_second": 32.806,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.9592437148094177,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6773,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.14084507042253522,
      "eval_loss": 0.8489218354225159,
      "eval_runtime": 0.0318,
      "eval_samples_per_second": 2230.929,
      "eval_steps_per_second": 31.422,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 99911832606720.0,
      "train_loss": 0.6877942879994711,
      "train_runtime": 6.9447,
      "train_samples_per_second": 4571.802,
      "train_steps_per_second": 21.599
    }
  ],
  "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": 99911832606720.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}