trainer_state.json

{
  "best_global_step": 6,
  "best_metric": 0.69921875,
  "best_model_checkpoint": "tiny_bert_km_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": 1.4377379417419434,
      "learning_rate": 4.9e-05,
      "loss": 0.7123,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.7005391120910645,
      "eval_runtime": 0.0298,
      "eval_samples_per_second": 2381.965,
      "eval_steps_per_second": 33.549,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.2045204639434814,
      "learning_rate": 4.8e-05,
      "loss": 0.7076,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5492957746478874,
      "eval_loss": 0.69921875,
      "eval_runtime": 0.0298,
      "eval_samples_per_second": 2384.348,
      "eval_steps_per_second": 33.582,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.8754438161849976,
      "learning_rate": 4.7e-05,
      "loss": 0.6978,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7239218354225159,
      "eval_runtime": 0.0296,
      "eval_samples_per_second": 2401.597,
      "eval_steps_per_second": 33.825,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6323876976966858,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6962,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.707691490650177,
      "eval_runtime": 0.0431,
      "eval_samples_per_second": 1648.513,
      "eval_steps_per_second": 23.218,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.996510922908783,
      "learning_rate": 4.5e-05,
      "loss": 0.6953,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5070422535211268,
      "eval_loss": 0.7035650014877319,
      "eval_runtime": 0.0608,
      "eval_samples_per_second": 1168.009,
      "eval_steps_per_second": 16.451,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.6087551712989807,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6953,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.4084507042253521,
      "eval_loss": 0.7111024260520935,
      "eval_runtime": 0.044,
      "eval_samples_per_second": 1611.998,
      "eval_steps_per_second": 22.704,
      "step": 18
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.7805089354515076,
      "learning_rate": 4.3e-05,
      "loss": 0.692,
      "step": 21
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.29577464788732394,
      "eval_loss": 0.719410240650177,
      "eval_runtime": 0.0308,
      "eval_samples_per_second": 2302.531,
      "eval_steps_per_second": 32.43,
      "step": 21
    },
    {
      "epoch": 7.0,
      "step": 21,
      "total_flos": 116563804707840.0,
      "train_loss": 0.6994877429235549,
      "train_runtime": 7.3204,
      "train_samples_per_second": 4337.188,
      "train_steps_per_second": 20.491
    }
  ],
  "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
}