trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.1600451469421387,
  "best_model_checkpoint": "tiny_bert_km_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": 3.770402193069458,
      "learning_rate": 4.9e-05,
      "loss": 3.6231,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.19396406848906156,
      "eval_loss": 2.1600451469421387,
      "eval_pearson": 0.20537418930332568,
      "eval_runtime": 0.4748,
      "eval_samples_per_second": 3159.294,
      "eval_spearmanr": 0.18255394767479743,
      "eval_steps_per_second": 12.637,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 6.684188365936279,
      "learning_rate": 4.8e-05,
      "loss": 2.0508,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.1810634046412402,
      "eval_loss": 2.4198853969573975,
      "eval_pearson": 0.1847198870941592,
      "eval_runtime": 0.5271,
      "eval_samples_per_second": 2845.493,
      "eval_spearmanr": 0.17740692218832121,
      "eval_steps_per_second": 11.382,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 5.661541938781738,
      "learning_rate": 4.7e-05,
      "loss": 1.9051,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.21541159126992399,
      "eval_loss": 2.338207960128784,
      "eval_pearson": 0.2217260864077553,
      "eval_runtime": 0.4942,
      "eval_samples_per_second": 3035.21,
      "eval_spearmanr": 0.20909709613209265,
      "eval_steps_per_second": 12.141,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 11.7733154296875,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.7517,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.25634348434597354,
      "eval_loss": 2.2891106605529785,
      "eval_pearson": 0.26020001612461785,
      "eval_runtime": 0.4734,
      "eval_samples_per_second": 3168.487,
      "eval_spearmanr": 0.2524869525673293,
      "eval_steps_per_second": 12.674,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 8.230708122253418,
      "learning_rate": 4.5e-05,
      "loss": 1.5284,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.27502552587763496,
      "eval_loss": 2.3171780109405518,
      "eval_pearson": 0.2781794019323813,
      "eval_runtime": 0.4742,
      "eval_samples_per_second": 3162.955,
      "eval_spearmanr": 0.27187164982288864,
      "eval_steps_per_second": 12.652,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 10.2731351852417,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.3355,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.2639794858629222,
      "eval_loss": 2.5106892585754395,
      "eval_pearson": 0.2681793879674212,
      "eval_runtime": 0.4619,
      "eval_samples_per_second": 3247.795,
      "eval_spearmanr": 0.25977958375842314,
      "eval_steps_per_second": 12.991,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 904528923227136.0,
      "train_loss": 2.0324475385140683,
      "train_runtime": 26.2914,
      "train_samples_per_second": 10933.235,
      "train_steps_per_second": 43.741
    }
  ],
  "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
}