trainer_state.json

{
  "best_metric": 0.930045871559633,
  "best_model_checkpoint": "./electra-base-discriminator/fine_tuned_models/checkpoint-10525",
  "epoch": 10.0,
  "global_step": 21050,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 1.8398562688161604e-05,
      "loss": 0.2442,
      "step": 2105
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9277522935779816,
      "eval_loss": 0.2364984154701233,
      "eval_runtime": 2.1955,
      "eval_samples_per_second": 397.168,
      "eval_steps_per_second": 49.646,
      "step": 2105
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.6354277945032536e-05,
      "loss": 0.1202,
      "step": 4210
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9254587155963303,
      "eval_loss": 0.25499287247657776,
      "eval_runtime": 2.1936,
      "eval_samples_per_second": 397.523,
      "eval_steps_per_second": 49.69,
      "step": 4210
    },
    {
      "epoch": 3.0,
      "learning_rate": 1.4309993201903469e-05,
      "loss": 0.0847,
      "step": 6315
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.926605504587156,
      "eval_loss": 0.25377511978149414,
      "eval_runtime": 2.192,
      "eval_samples_per_second": 397.808,
      "eval_steps_per_second": 49.726,
      "step": 6315
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.2265708458774402e-05,
      "loss": 0.0645,
      "step": 8420
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9220183486238532,
      "eval_loss": 0.3068634867668152,
      "eval_runtime": 2.1868,
      "eval_samples_per_second": 398.761,
      "eval_steps_per_second": 49.845,
      "step": 8420
    },
    {
      "epoch": 5.0,
      "learning_rate": 1.0221423715645335e-05,
      "loss": 0.046,
      "step": 10525
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.930045871559633,
      "eval_loss": 0.30590787529945374,
      "eval_runtime": 2.1975,
      "eval_samples_per_second": 396.806,
      "eval_steps_per_second": 49.601,
      "step": 10525
    },
    {
      "epoch": 6.0,
      "learning_rate": 8.177138972516268e-06,
      "loss": 0.0325,
      "step": 12630
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9231651376146789,
      "eval_loss": 0.36431288719177246,
      "eval_runtime": 2.202,
      "eval_samples_per_second": 396.006,
      "eval_steps_per_second": 49.501,
      "step": 12630
    },
    {
      "epoch": 7.0,
      "learning_rate": 6.132854229387201e-06,
      "loss": 0.024,
      "step": 14735
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9208715596330275,
      "eval_loss": 0.42286497354507446,
      "eval_runtime": 2.2039,
      "eval_samples_per_second": 395.666,
      "eval_steps_per_second": 49.458,
      "step": 14735
    },
    {
      "epoch": 8.0,
      "learning_rate": 4.088569486258134e-06,
      "loss": 0.0176,
      "step": 16840
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.926605504587156,
      "eval_loss": 0.45770832896232605,
      "eval_runtime": 2.22,
      "eval_samples_per_second": 392.794,
      "eval_steps_per_second": 49.099,
      "step": 16840
    },
    {
      "epoch": 9.0,
      "learning_rate": 2.044284743129067e-06,
      "loss": 0.0111,
      "step": 18945
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9254587155963303,
      "eval_loss": 0.5075973868370056,
      "eval_runtime": 2.1989,
      "eval_samples_per_second": 396.566,
      "eval_steps_per_second": 49.571,
      "step": 18945
    },
    {
      "epoch": 10.0,
      "learning_rate": 0.0,
      "loss": 0.0079,
      "step": 21050
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9208715596330275,
      "eval_loss": 0.585883378982544,
      "eval_runtime": 2.2147,
      "eval_samples_per_second": 393.74,
      "eval_steps_per_second": 49.217,
      "step": 21050
    },
    {
      "epoch": 10.0,
      "step": 21050,
      "total_flos": 2.831471218376448e+16,
      "train_loss": 0.06527320399703436,
      "train_runtime": 4410.9322,
      "train_samples_per_second": 152.687,
      "train_steps_per_second": 4.772
    }
  ],
  "max_steps": 21050,
  "num_train_epochs": 10,
  "total_flos": 2.831471218376448e+16,
  "trial_name": null,
  "trial_params": null
}