trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.3185579776763916,
  "best_model_checkpoint": "bert_base_rand_20_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": 4.732888698577881,
      "learning_rate": 4.9e-05,
      "loss": 2.5033,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.1410360200001674,
      "eval_loss": 2.3185579776763916,
      "eval_pearson": 0.15053041966314645,
      "eval_runtime": 0.9535,
      "eval_samples_per_second": 1573.101,
      "eval_spearmanr": 0.13154162033718833,
      "eval_steps_per_second": 6.292,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.306790351867676,
      "learning_rate": 4.8e-05,
      "loss": 1.8985,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.18015705657848313,
      "eval_loss": 2.529568672180176,
      "eval_pearson": 0.18550678822396657,
      "eval_runtime": 0.964,
      "eval_samples_per_second": 1556.054,
      "eval_spearmanr": 0.17480732493299966,
      "eval_steps_per_second": 6.224,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 6.404316425323486,
      "learning_rate": 4.7e-05,
      "loss": 1.6671,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.20184561161741704,
      "eval_loss": 2.597043037414551,
      "eval_pearson": 0.20188316890864003,
      "eval_runtime": 0.9762,
      "eval_samples_per_second": 1536.637,
      "eval_spearmanr": 0.20180805432619406,
      "eval_steps_per_second": 6.147,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 12.172585487365723,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.3208,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.2953607717701988,
      "eval_loss": 2.3513340950012207,
      "eval_pearson": 0.2942727905500899,
      "eval_runtime": 0.9915,
      "eval_samples_per_second": 1512.802,
      "eval_spearmanr": 0.29644875299030776,
      "eval_steps_per_second": 6.051,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 13.733288764953613,
      "learning_rate": 4.5e-05,
      "loss": 0.982,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.2777316128111841,
      "eval_loss": 2.5606610774993896,
      "eval_pearson": 0.27991713875292457,
      "eval_runtime": 0.949,
      "eval_samples_per_second": 1580.659,
      "eval_spearmanr": 0.2755460868694436,
      "eval_steps_per_second": 6.323,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 10.061882019042969,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.7114,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.3250467045437194,
      "eval_loss": 2.4146175384521484,
      "eval_pearson": 0.3261100716445622,
      "eval_runtime": 0.9748,
      "eval_samples_per_second": 1538.765,
      "eval_spearmanr": 0.3239833374428766,
      "eval_steps_per_second": 6.155,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 4537835628033024.0,
      "train_loss": 1.5138361419456592,
      "train_runtime": 74.5311,
      "train_samples_per_second": 3856.78,
      "train_steps_per_second": 15.43
    }
  ],
  "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": 4537835628033024.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}