trainer_state.json

{
  "best_metric": 0.2546990215778351,
  "best_model_checkpoint": "xlm-roberta-base-all-finetuned-toxicity-classification/checkpoint-2532",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 7596,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "learning_rate": 1.8683517640863615e-05,
      "loss": 0.2799,
      "step": 500
    },
    {
      "epoch": 0.39,
      "learning_rate": 1.7367035281727228e-05,
      "loss": 0.2876,
      "step": 1000
    },
    {
      "epoch": 0.59,
      "learning_rate": 1.6050552922590838e-05,
      "loss": 0.2671,
      "step": 1500
    },
    {
      "epoch": 0.79,
      "learning_rate": 1.473407056345445e-05,
      "loss": 0.2639,
      "step": 2000
    },
    {
      "epoch": 0.99,
      "learning_rate": 1.3417588204318064e-05,
      "loss": 0.2675,
      "step": 2500
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.2546990215778351,
      "eval_runtime": 35.0277,
      "eval_samples_per_second": 128.47,
      "eval_steps_per_second": 8.051,
      "step": 2532
    },
    {
      "epoch": 1.18,
      "learning_rate": 1.2101105845181676e-05,
      "loss": 0.2072,
      "step": 3000
    },
    {
      "epoch": 1.38,
      "learning_rate": 1.0784623486045287e-05,
      "loss": 0.2035,
      "step": 3500
    },
    {
      "epoch": 1.58,
      "learning_rate": 9.4681411269089e-06,
      "loss": 0.1987,
      "step": 4000
    },
    {
      "epoch": 1.78,
      "learning_rate": 8.151658767772512e-06,
      "loss": 0.1938,
      "step": 4500
    },
    {
      "epoch": 1.97,
      "learning_rate": 6.835176408636125e-06,
      "loss": 0.189,
      "step": 5000
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.29669827222824097,
      "eval_runtime": 35.0287,
      "eval_samples_per_second": 128.466,
      "eval_steps_per_second": 8.051,
      "step": 5064
    },
    {
      "epoch": 2.17,
      "learning_rate": 5.518694049499738e-06,
      "loss": 0.1412,
      "step": 5500
    },
    {
      "epoch": 2.37,
      "learning_rate": 4.20221169036335e-06,
      "loss": 0.1467,
      "step": 6000
    },
    {
      "epoch": 2.57,
      "learning_rate": 2.885729331226962e-06,
      "loss": 0.1397,
      "step": 6500
    },
    {
      "epoch": 2.76,
      "learning_rate": 1.5692469720905742e-06,
      "loss": 0.1476,
      "step": 7000
    },
    {
      "epoch": 2.96,
      "learning_rate": 2.527646129541864e-07,
      "loss": 0.1383,
      "step": 7500
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.4095066785812378,
      "eval_runtime": 35.0202,
      "eval_samples_per_second": 128.497,
      "eval_steps_per_second": 8.053,
      "step": 7596
    }
  ],
  "logging_steps": 500,
  "max_steps": 7596,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3.196799322624e+16,
  "trial_name": null,
  "trial_params": null
}