trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.3236286640167236,
  "best_model_checkpoint": "tiny_bert_rand_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": 2.5193071365356445,
      "learning_rate": 4.9e-05,
      "loss": 2.9227,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.07619254605764625,
      "eval_loss": 2.3236286640167236,
      "eval_pearson": 0.07898484684438345,
      "eval_runtime": 0.454,
      "eval_samples_per_second": 3303.958,
      "eval_spearmanr": 0.07340024527090903,
      "eval_steps_per_second": 13.216,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 5.809595584869385,
      "learning_rate": 4.8e-05,
      "loss": 1.9914,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.08750659038872105,
      "eval_loss": 2.634969711303711,
      "eval_pearson": 0.09812738467941336,
      "eval_runtime": 0.506,
      "eval_samples_per_second": 2964.196,
      "eval_spearmanr": 0.07688579609802873,
      "eval_steps_per_second": 11.857,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.328521251678467,
      "learning_rate": 4.7e-05,
      "loss": 1.8134,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.1859433210745351,
      "eval_loss": 2.440565347671509,
      "eval_pearson": 0.18896606102458452,
      "eval_runtime": 0.4676,
      "eval_samples_per_second": 3207.972,
      "eval_spearmanr": 0.18292058112448567,
      "eval_steps_per_second": 12.832,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.402108192443848,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.6045,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.23259206775338748,
      "eval_loss": 2.4218363761901855,
      "eval_pearson": 0.2332777049263922,
      "eval_runtime": 0.4521,
      "eval_samples_per_second": 3318.166,
      "eval_spearmanr": 0.23190643058038274,
      "eval_steps_per_second": 13.273,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 9.683623313903809,
      "learning_rate": 4.5e-05,
      "loss": 1.3303,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.2551059052757607,
      "eval_loss": 2.545163154602051,
      "eval_pearson": 0.2552566794086755,
      "eval_runtime": 0.4911,
      "eval_samples_per_second": 3054.412,
      "eval_spearmanr": 0.25495513114284585,
      "eval_steps_per_second": 12.218,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 7.816468238830566,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 1.1276,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.267269762918755,
      "eval_loss": 2.635294198989868,
      "eval_pearson": 0.2657969708338703,
      "eval_runtime": 0.4578,
      "eval_samples_per_second": 3276.496,
      "eval_spearmanr": 0.26874255500363964,
      "eval_steps_per_second": 13.106,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 904528923227136.0,
      "train_loss": 1.7983283720154692,
      "train_runtime": 26.4474,
      "train_samples_per_second": 10868.76,
      "train_steps_per_second": 43.483
    }
  ],
  "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
}