trainer_state.json

{
  "best_global_step": 30,
  "best_metric": 0.6790599226951599,
  "best_model_checkpoint": "tinybert_base_train_kd_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": 0.5859740972518921,
      "learning_rate": 4.91e-05,
      "loss": 0.705,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5270758122743683,
      "eval_loss": 0.691801130771637,
      "eval_runtime": 0.1452,
      "eval_samples_per_second": 1907.112,
      "eval_steps_per_second": 13.77,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.07582426071167,
      "learning_rate": 4.8100000000000004e-05,
      "loss": 0.695,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.48014440433212996,
      "eval_loss": 0.7236151695251465,
      "eval_runtime": 0.1475,
      "eval_samples_per_second": 1877.96,
      "eval_steps_per_second": 13.559,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.8762505054473877,
      "learning_rate": 4.71e-05,
      "loss": 0.6831,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5848375451263538,
      "eval_loss": 0.6790599226951599,
      "eval_runtime": 0.1418,
      "eval_samples_per_second": 1953.151,
      "eval_steps_per_second": 14.102,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 3.220576286315918,
      "learning_rate": 4.61e-05,
      "loss": 0.6248,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.5884476534296029,
      "eval_loss": 0.6915948390960693,
      "eval_runtime": 0.1392,
      "eval_samples_per_second": 1989.957,
      "eval_steps_per_second": 14.368,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 5.529593467712402,
      "learning_rate": 4.5100000000000005e-05,
      "loss": 0.5451,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.6245487364620939,
      "eval_loss": 0.7247257232666016,
      "eval_runtime": 0.1471,
      "eval_samples_per_second": 1883.162,
      "eval_steps_per_second": 13.597,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 4.2390875816345215,
      "learning_rate": 4.41e-05,
      "loss": 0.4932,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5884476534296029,
      "eval_loss": 0.7858375310897827,
      "eval_runtime": 0.1629,
      "eval_samples_per_second": 1700.836,
      "eval_steps_per_second": 12.28,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 5.121857643127441,
      "learning_rate": 4.3100000000000004e-05,
      "loss": 0.3597,
      "step": 70
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6028880866425993,
      "eval_loss": 0.9329966306686401,
      "eval_runtime": 0.1713,
      "eval_samples_per_second": 1617.04,
      "eval_steps_per_second": 11.675,
      "step": 70
    },
    {
      "epoch": 8.0,
      "grad_norm": 3.846261501312256,
      "learning_rate": 4.21e-05,
      "loss": 0.268,
      "step": 80
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.6209386281588448,
      "eval_loss": 0.9830819964408875,
      "eval_runtime": 0.1514,
      "eval_samples_per_second": 1829.592,
      "eval_steps_per_second": 13.21,
      "step": 80
    },
    {
      "epoch": 8.0,
      "step": 80,
      "total_flos": 1319375290613760.0,
      "train_loss": 0.5467315256595612,
      "train_runtime": 34.017,
      "train_samples_per_second": 3659.931,
      "train_steps_per_second": 14.699
    }
  ],
  "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": 1319375290613760.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}