trainer_state.json

{
  "best_metric": 0.6930568218231201,
  "best_model_checkpoint": "tiny_bert_km_100_v1_wnli/checkpoint-3",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.8802371621131897,
      "learning_rate": 4.9e-05,
      "loss": 0.7061,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5211267605633803,
      "eval_loss": 0.6930568218231201,
      "eval_runtime": 0.0372,
      "eval_samples_per_second": 1907.54,
      "eval_steps_per_second": 26.867,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.7727060914039612,
      "learning_rate": 4.8e-05,
      "loss": 0.6931,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.49295774647887325,
      "eval_loss": 0.6975681781768799,
      "eval_runtime": 0.0365,
      "eval_samples_per_second": 1946.16,
      "eval_steps_per_second": 27.411,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6414340138435364,
      "learning_rate": 4.7e-05,
      "loss": 0.6991,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.38028169014084506,
      "eval_loss": 0.7135781645774841,
      "eval_runtime": 0.0381,
      "eval_samples_per_second": 1861.106,
      "eval_steps_per_second": 26.213,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.7113948464393616,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.692,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.38028169014084506,
      "eval_loss": 0.719905436038971,
      "eval_runtime": 0.0377,
      "eval_samples_per_second": 1882.983,
      "eval_steps_per_second": 26.521,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.803973913192749,
      "learning_rate": 4.5e-05,
      "loss": 0.6907,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.36619718309859156,
      "eval_loss": 0.7242516279220581,
      "eval_runtime": 0.0386,
      "eval_samples_per_second": 1841.427,
      "eval_steps_per_second": 25.936,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.8193577527999878,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6884,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.38028169014084506,
      "eval_loss": 0.7314040064811707,
      "eval_runtime": 0.0412,
      "eval_samples_per_second": 1722.527,
      "eval_steps_per_second": 24.261,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 99911832606720.0,
      "train_loss": 0.6949052943123711,
      "train_runtime": 8.1637,
      "train_samples_per_second": 3889.17,
      "train_steps_per_second": 18.374
    }
  ],
  "logging_steps": 1,
  "max_steps": 150,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "total_flos": 99911832606720.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}