trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.2269647121429443,
  "best_model_checkpoint": "tiny_bert_km_10_v1_stsb/checkpoint-23",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 138,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.8133530616760254,
      "learning_rate": 4.9e-05,
      "loss": 3.5394,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.13555081068568664,
      "eval_loss": 2.2269647121429443,
      "eval_pearson": 0.1335095420226322,
      "eval_runtime": 0.4461,
      "eval_samples_per_second": 3362.568,
      "eval_spearmanr": 0.13759207934874104,
      "eval_steps_per_second": 13.45,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 9.366716384887695,
      "learning_rate": 4.8e-05,
      "loss": 2.1228,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.16277143083477807,
      "eval_loss": 2.285903215408325,
      "eval_pearson": 0.15043910557772164,
      "eval_runtime": 0.4772,
      "eval_samples_per_second": 3143.105,
      "eval_spearmanr": 0.17510375609183448,
      "eval_steps_per_second": 12.572,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 3.574167013168335,
      "learning_rate": 4.7e-05,
      "loss": 1.9666,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.18471718213227462,
      "eval_loss": 2.3973381519317627,
      "eval_pearson": 0.18366938050701523,
      "eval_runtime": 0.486,
      "eval_samples_per_second": 3086.48,
      "eval_spearmanr": 0.18576498375753403,
      "eval_steps_per_second": 12.346,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.124925136566162,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.849,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.18076252838733606,
      "eval_loss": 2.5342788696289062,
      "eval_pearson": 0.18941767035860568,
      "eval_runtime": 0.4356,
      "eval_samples_per_second": 3443.326,
      "eval_spearmanr": 0.17210738641606643,
      "eval_steps_per_second": 13.773,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 9.790884971618652,
      "learning_rate": 4.5e-05,
      "loss": 1.6195,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.22900800322858658,
      "eval_loss": 2.620011329650879,
      "eval_pearson": 0.2320085793770834,
      "eval_runtime": 0.4417,
      "eval_samples_per_second": 3395.95,
      "eval_spearmanr": 0.22600742708008975,
      "eval_steps_per_second": 13.584,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 6.0991129875183105,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.4318,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.2485673987468651,
      "eval_loss": 2.562162160873413,
      "eval_pearson": 0.2510206065778135,
      "eval_runtime": 0.4672,
      "eval_samples_per_second": 3210.498,
      "eval_spearmanr": 0.2461141909159167,
      "eval_steps_per_second": 12.842,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 904528923227136.0,
      "train_loss": 2.088208903437075,
      "train_runtime": 25.4978,
      "train_samples_per_second": 11273.536,
      "train_steps_per_second": 45.102
    }
  ],
  "logging_steps": 1,
  "max_steps": 1150,
  "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": 904528923227136.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}