trainer_state.json

{
  "best_global_step": 3,
  "best_metric": 0.713358461856842,
  "best_model_checkpoint": "bert_base_rand_5_v2_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": 1.9324321746826172,
      "learning_rate": 4.9e-05,
      "loss": 0.7423,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5352112676056338,
      "eval_loss": 0.713358461856842,
      "eval_runtime": 0.0557,
      "eval_samples_per_second": 1273.589,
      "eval_steps_per_second": 17.938,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7285646200180054,
      "learning_rate": 4.8e-05,
      "loss": 0.695,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7621039152145386,
      "eval_runtime": 0.0574,
      "eval_samples_per_second": 1236.487,
      "eval_steps_per_second": 17.415,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.7954620122909546,
      "learning_rate": 4.7e-05,
      "loss": 0.6971,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5070422535211268,
      "eval_loss": 0.7425177097320557,
      "eval_runtime": 0.0574,
      "eval_samples_per_second": 1236.451,
      "eval_steps_per_second": 17.415,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.9232250452041626,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.7029,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.23943661971830985,
      "eval_loss": 0.7899428606033325,
      "eval_runtime": 0.058,
      "eval_samples_per_second": 1224.564,
      "eval_steps_per_second": 17.247,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8401814103126526,
      "learning_rate": 4.5e-05,
      "loss": 0.6869,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.19718309859154928,
      "eval_loss": 0.8493890762329102,
      "eval_runtime": 0.0685,
      "eval_samples_per_second": 1036.828,
      "eval_steps_per_second": 14.603,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 1.2261316776275635,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6901,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.15492957746478872,
      "eval_loss": 0.8630609512329102,
      "eval_runtime": 0.0555,
      "eval_samples_per_second": 1278.63,
      "eval_steps_per_second": 18.009,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 501226560460800.0,
      "train_loss": 0.7023815976248847,
      "train_runtime": 19.5957,
      "train_samples_per_second": 1620.254,
      "train_steps_per_second": 7.655
    }
  ],
  "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": 501226560460800.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}