models/bert-offensive/checkpoint-3500/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.348993288590604,
  "eval_steps": 500,
  "global_step": 3500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.34,
      "grad_norm": 3.309788227081299,
      "learning_rate": 4.440715883668904e-05,
      "loss": 0.5291,
      "step": 500
    },
    {
      "epoch": 0.67,
      "grad_norm": 8.38113784790039,
      "learning_rate": 3.881431767337807e-05,
      "loss": 0.4834,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8244186046511628,
      "eval_f1": 0.700990099009901,
      "eval_loss": 0.40849873423576355,
      "eval_precision": 0.6679245283018868,
      "eval_recall": 0.7375,
      "eval_runtime": 1.415,
      "eval_samples_per_second": 607.771,
      "eval_steps_per_second": 76.325,
      "step": 1490
    },
    {
      "epoch": 1.01,
      "grad_norm": 4.857753753662109,
      "learning_rate": 3.3221476510067115e-05,
      "loss": 0.4743,
      "step": 1500
    },
    {
      "epoch": 1.34,
      "grad_norm": 20.212223052978516,
      "learning_rate": 2.7628635346756153e-05,
      "loss": 0.3753,
      "step": 2000
    },
    {
      "epoch": 1.68,
      "grad_norm": 2.194687604904175,
      "learning_rate": 2.2035794183445192e-05,
      "loss": 0.3819,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8383720930232558,
      "eval_f1": 0.7010752688172043,
      "eval_loss": 0.46891969442367554,
      "eval_precision": 0.7244444444444444,
      "eval_recall": 0.6791666666666667,
      "eval_runtime": 1.4131,
      "eval_samples_per_second": 608.584,
      "eval_steps_per_second": 76.427,
      "step": 2980
    },
    {
      "epoch": 2.01,
      "grad_norm": 1.1423423290252686,
      "learning_rate": 1.644295302013423e-05,
      "loss": 0.3561,
      "step": 3000
    },
    {
      "epoch": 2.35,
      "grad_norm": 112.02091979980469,
      "learning_rate": 1.0850111856823267e-05,
      "loss": 0.2597,
      "step": 3500
    }
  ],
  "logging_steps": 500,
  "max_steps": 4470,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1481335797228000.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}