trainer_state.json

{
  "best_global_step": 23,
  "best_metric": 2.308342695236206,
  "best_model_checkpoint": "distilbert_km_10_v2_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": 4.890619277954102,
      "learning_rate": 4.9e-05,
      "loss": 3.0233,
      "step": 23
    },
    {
      "epoch": 1.0,
      "eval_combined_score": 0.14546301722281513,
      "eval_loss": 2.308342695236206,
      "eval_pearson": 0.1514607194595729,
      "eval_runtime": 0.6274,
      "eval_samples_per_second": 2390.705,
      "eval_spearmanr": 0.13946531498605738,
      "eval_steps_per_second": 9.563,
      "step": 23
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.947163105010986,
      "learning_rate": 4.8e-05,
      "loss": 1.9441,
      "step": 46
    },
    {
      "epoch": 2.0,
      "eval_combined_score": 0.1680556586508381,
      "eval_loss": 2.496310234069824,
      "eval_pearson": 0.17750219984139104,
      "eval_runtime": 0.6282,
      "eval_samples_per_second": 2387.816,
      "eval_spearmanr": 0.15860911746028514,
      "eval_steps_per_second": 9.551,
      "step": 46
    },
    {
      "epoch": 3.0,
      "grad_norm": 7.757569789886475,
      "learning_rate": 4.7e-05,
      "loss": 1.6958,
      "step": 69
    },
    {
      "epoch": 3.0,
      "eval_combined_score": 0.21327894535996159,
      "eval_loss": 2.3428003787994385,
      "eval_pearson": 0.22355298733190435,
      "eval_runtime": 0.668,
      "eval_samples_per_second": 2245.459,
      "eval_spearmanr": 0.20300490338801883,
      "eval_steps_per_second": 8.982,
      "step": 69
    },
    {
      "epoch": 4.0,
      "grad_norm": 6.711005210876465,
      "learning_rate": 4.600000000000001e-05,
      "loss": 1.428,
      "step": 92
    },
    {
      "epoch": 4.0,
      "eval_combined_score": 0.2606544358399977,
      "eval_loss": 2.5911128520965576,
      "eval_pearson": 0.2649052185675204,
      "eval_runtime": 0.622,
      "eval_samples_per_second": 2411.553,
      "eval_spearmanr": 0.25640365311247504,
      "eval_steps_per_second": 9.646,
      "step": 92
    },
    {
      "epoch": 5.0,
      "grad_norm": 9.803565979003906,
      "learning_rate": 4.5e-05,
      "loss": 1.1342,
      "step": 115
    },
    {
      "epoch": 5.0,
      "eval_combined_score": 0.2964874473964536,
      "eval_loss": 2.484288215637207,
      "eval_pearson": 0.29901417650483925,
      "eval_runtime": 0.6958,
      "eval_samples_per_second": 2155.696,
      "eval_spearmanr": 0.29396071828806797,
      "eval_steps_per_second": 8.623,
      "step": 115
    },
    {
      "epoch": 6.0,
      "grad_norm": 8.161381721496582,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.828,
      "step": 138
    },
    {
      "epoch": 6.0,
      "eval_combined_score": 0.31314439435783836,
      "eval_loss": 2.4830892086029053,
      "eval_pearson": 0.3131639464327934,
      "eval_runtime": 0.733,
      "eval_samples_per_second": 2046.29,
      "eval_spearmanr": 0.31312484228288334,
      "eval_steps_per_second": 8.185,
      "step": 138
    },
    {
      "epoch": 6.0,
      "step": 138,
      "total_flos": 2284624480859136.0,
      "train_loss": 1.6755805913952813,
      "train_runtime": 44.153,
      "train_samples_per_second": 6510.321,
      "train_steps_per_second": 26.046
    }
  ],
  "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": 2284624480859136.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}