trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.2005562782287598,
  "best_model_checkpoint": "tiny_bert_km_5_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": 7.434397220611572,
      "learning_rate": 4.9e-05,
      "loss": 3.4235,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.1465144909713814,
      "eval_loss": 2.2005562782287598,
      "eval_pearson": 0.16116672514687252,
      "eval_runtime": 0.5131,
      "eval_samples_per_second": 2923.334,
      "eval_spearmanr": 0.13186225679589025,
      "eval_steps_per_second": 11.693,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.577157497406006,
      "learning_rate": 4.8e-05,
      "loss": 2.0135,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.16466816617192903,
      "eval_loss": 2.5477442741394043,
      "eval_pearson": 0.16970656234839473,
      "eval_runtime": 0.4526,
      "eval_samples_per_second": 3314.169,
      "eval_spearmanr": 0.1596297699954633,
      "eval_steps_per_second": 13.257,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 8.487744331359863,
      "learning_rate": 4.7e-05,
      "loss": 1.805,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.2604221859382768,
      "eval_loss": 2.309962511062622,
      "eval_pearson": 0.2644087582008461,
      "eval_runtime": 0.4744,
      "eval_samples_per_second": 3162.085,
      "eval_spearmanr": 0.2564356136757076,
      "eval_steps_per_second": 12.648,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 7.679525375366211,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.566,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.28294468742144885,
      "eval_loss": 2.4052906036376953,
      "eval_pearson": 0.28411820655844633,
      "eval_runtime": 0.4569,
      "eval_samples_per_second": 3283.101,
      "eval_spearmanr": 0.2817711682844513,
      "eval_steps_per_second": 13.132,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 16.149778366088867,
      "learning_rate": 4.5e-05,
      "loss": 1.3172,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.28830085583325116,
      "eval_loss": 2.4673807621002197,
      "eval_pearson": 0.2893346672156352,
      "eval_runtime": 0.4524,
      "eval_samples_per_second": 3315.86,
      "eval_spearmanr": 0.28726704445086704,
      "eval_steps_per_second": 13.263,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 13.063384056091309,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.1328,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.3164627804210498,
      "eval_loss": 2.3672008514404297,
      "eval_pearson": 0.31793862253537136,
      "eval_runtime": 0.4874,
      "eval_samples_per_second": 3077.292,
      "eval_spearmanr": 0.3149869383067282,
      "eval_steps_per_second": 12.309,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 904528923227136.0,
      "train_loss": 1.8763193600419639,
      "train_runtime": 26.7431,
      "train_samples_per_second": 10748.571,
      "train_steps_per_second": 43.002
    }
  ],
  "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
}