trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.6836854219436646,
  "best_model_checkpoint": "tiny_bert_rand_50_v2_rte/checkpoint-30",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 80,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.41052356362342834,
      "learning_rate": 4.9e-05,
      "loss": 0.701,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.555956678700361,
      "eval_loss": 0.6892063617706299,
      "eval_runtime": 0.098,
      "eval_samples_per_second": 2825.093,
      "eval_steps_per_second": 20.398,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.4319133460521698,
      "learning_rate": 4.8e-05,
      "loss": 0.6924,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5667870036101083,
      "eval_loss": 0.6860898733139038,
      "eval_runtime": 0.1117,
      "eval_samples_per_second": 2479.67,
      "eval_steps_per_second": 17.904,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.551663875579834,
      "learning_rate": 4.7e-05,
      "loss": 0.6817,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5776173285198556,
      "eval_loss": 0.6836854219436646,
      "eval_runtime": 0.109,
      "eval_samples_per_second": 2540.729,
      "eval_steps_per_second": 18.345,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.0392649173736572,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6475,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5812274368231047,
      "eval_loss": 0.7117906808853149,
      "eval_runtime": 0.1119,
      "eval_samples_per_second": 2475.154,
      "eval_steps_per_second": 17.871,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.1567509174346924,
      "learning_rate": 4.5e-05,
      "loss": 0.5913,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5379061371841155,
      "eval_loss": 0.7820150256156921,
      "eval_runtime": 0.1043,
      "eval_samples_per_second": 2654.72,
      "eval_steps_per_second": 19.168,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.476806640625,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5074,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.555956678700361,
      "eval_loss": 0.8313120603561401,
      "eval_runtime": 0.1216,
      "eval_samples_per_second": 2277.502,
      "eval_steps_per_second": 16.444,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 1.9686747789382935,
      "learning_rate": 4.3e-05,
      "loss": 0.3845,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.5415162454873647,
      "eval_loss": 1.0666288137435913,
      "eval_runtime": 0.1016,
      "eval_samples_per_second": 2727.237,
      "eval_steps_per_second": 19.691,
      "step": 70
    },
    {
      "epoch": 8.0,
      "grad_norm": 3.5235719680786133,
      "learning_rate": 4.2e-05,
      "loss": 0.2643,
      "step": 80
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.5415162454873647,
      "eval_loss": 1.234315037727356,
      "eval_runtime": 0.0998,
      "eval_samples_per_second": 2775.468,
      "eval_steps_per_second": 20.039,
      "step": 80
    },
    {
      "epoch": 8.0,
      "step": 80,
      "total_flos": 522373675991040.0,
      "train_loss": 0.5587728291749954,
      "train_runtime": 18.0236,
      "train_samples_per_second": 6907.6,
      "train_steps_per_second": 27.741
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "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": 522373675991040.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}