trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.70703125,
  "best_model_checkpoint": "tiny_bert_rand_5_v1_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": 1.1953152418136597,
      "learning_rate": 4.9e-05,
      "loss": 0.7071,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.39436619718309857,
      "eval_loss": 0.710222065448761,
      "eval_runtime": 0.0322,
      "eval_samples_per_second": 2204.995,
      "eval_steps_per_second": 31.056,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.577671229839325,
      "learning_rate": 4.8e-05,
      "loss": 0.6958,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5070422535211268,
      "eval_loss": 0.70703125,
      "eval_runtime": 0.0302,
      "eval_samples_per_second": 2347.749,
      "eval_steps_per_second": 33.067,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6530124545097351,
      "learning_rate": 4.7e-05,
      "loss": 0.6966,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.3380281690140845,
      "eval_loss": 0.719410240650177,
      "eval_runtime": 0.0335,
      "eval_samples_per_second": 2117.838,
      "eval_steps_per_second": 29.829,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.7644088864326477,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6893,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.22535211267605634,
      "eval_loss": 0.731624186038971,
      "eval_runtime": 0.0324,
      "eval_samples_per_second": 2191.623,
      "eval_steps_per_second": 30.868,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.6378266215324402,
      "learning_rate": 4.5e-05,
      "loss": 0.6887,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.22535211267605634,
      "eval_loss": 0.740702211856842,
      "eval_runtime": 0.0337,
      "eval_samples_per_second": 2103.76,
      "eval_steps_per_second": 29.63,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.7794684171676636,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6904,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.22535211267605634,
      "eval_loss": 0.7487897872924805,
      "eval_runtime": 0.0321,
      "eval_samples_per_second": 2208.429,
      "eval_steps_per_second": 31.105,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.5502309799194336,
      "learning_rate": 4.3e-05,
      "loss": 0.6899,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.28169014084507044,
      "eval_loss": 0.756657063961029,
      "eval_runtime": 0.031,
      "eval_samples_per_second": 2287.955,
      "eval_steps_per_second": 32.225,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 116563804707840.0,
      "train_loss": 0.693983827318464,
      "train_runtime": 7.816,
      "train_samples_per_second": 4062.194,
      "train_steps_per_second": 19.191
    }
  ],
  "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
}