trainer_state.json

{
  "best_global_step": 820,
  "best_metric": 0.6248087882995605,
  "best_model_checkpoint": "bert_base_km_50_v2_qnli/checkpoint-820",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 2870,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.7058974504470825,
      "learning_rate": 4.9e-05,
      "loss": 0.6636,
      "step": 410
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6209042650558302,
      "eval_loss": 0.6401111483573914,
      "eval_runtime": 3.4997,
      "eval_samples_per_second": 1560.972,
      "eval_steps_per_second": 6.286,
      "step": 410
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7402713298797607,
      "learning_rate": 4.8e-05,
      "loss": 0.6252,
      "step": 820
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.6512904997254256,
      "eval_loss": 0.6248087882995605,
      "eval_runtime": 3.481,
      "eval_samples_per_second": 1569.355,
      "eval_steps_per_second": 6.32,
      "step": 820
    },
    {
      "epoch": 3.0,
      "grad_norm": 4.2072062492370605,
      "learning_rate": 4.7e-05,
      "loss": 0.5582,
      "step": 1230
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6360973823906279,
      "eval_loss": 0.6467459797859192,
      "eval_runtime": 3.4617,
      "eval_samples_per_second": 1578.12,
      "eval_steps_per_second": 6.355,
      "step": 1230
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.934783935546875,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.4467,
      "step": 1640
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.6560497894929526,
      "eval_loss": 0.7168631553649902,
      "eval_runtime": 3.439,
      "eval_samples_per_second": 1588.549,
      "eval_steps_per_second": 6.397,
      "step": 1640
    },
    {
      "epoch": 5.0,
      "grad_norm": 6.266510963439941,
      "learning_rate": 4.5e-05,
      "loss": 0.3229,
      "step": 2050
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.651473549331869,
      "eval_loss": 0.8077744841575623,
      "eval_runtime": 3.4441,
      "eval_samples_per_second": 1586.195,
      "eval_steps_per_second": 6.388,
      "step": 2050
    },
    {
      "epoch": 6.0,
      "grad_norm": 9.762896537780762,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.2249,
      "step": 2460
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.647263408383672,
      "eval_loss": 0.9613845944404602,
      "eval_runtime": 3.4507,
      "eval_samples_per_second": 1583.144,
      "eval_steps_per_second": 6.375,
      "step": 2460
    },
    {
      "epoch": 7.0,
      "grad_norm": 13.420165061950684,
      "learning_rate": 4.3e-05,
      "loss": 0.1597,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.6450668131063518,
      "eval_loss": 1.1662894487380981,
      "eval_runtime": 3.4427,
      "eval_samples_per_second": 1586.817,
      "eval_steps_per_second": 6.39,
      "step": 2870
    },
    {
      "epoch": 7.0,
      "step": 2870,
      "total_flos": 9.645664445050368e+16,
      "train_loss": 0.428730151877586,
      "train_runtime": 1213.5853,
      "train_samples_per_second": 4315.436,
      "train_steps_per_second": 16.892
    }
  ],
  "logging_steps": 1,
  "max_steps": 20500,
  "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": 9.645664445050368e+16,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}