trainer_state.json

{
  "best_metric": 0.69140625,
  "best_model_checkpoint": "tiny_bert_rand_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": 0.5936211347579956,
      "learning_rate": 4.9e-05,
      "loss": 0.6941,
      "step": 3
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.69140625,
      "eval_runtime": 0.0465,
      "eval_samples_per_second": 1527.008,
      "eval_steps_per_second": 21.507,
      "step": 3
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.4567447602748871,
      "learning_rate": 4.8e-05,
      "loss": 0.6962,
      "step": 6
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5633802816901409,
      "eval_loss": 0.6954226493835449,
      "eval_runtime": 0.0398,
      "eval_samples_per_second": 1783.41,
      "eval_steps_per_second": 25.118,
      "step": 6
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.458055317401886,
      "learning_rate": 4.7e-05,
      "loss": 0.6966,
      "step": 9
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.43661971830985913,
      "eval_loss": 0.7034552097320557,
      "eval_runtime": 0.0365,
      "eval_samples_per_second": 1943.835,
      "eval_steps_per_second": 27.378,
      "step": 9
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.6440699100494385,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6913,
      "step": 12
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.4084507042253521,
      "eval_loss": 0.6997138857841492,
      "eval_runtime": 0.0352,
      "eval_samples_per_second": 2018.447,
      "eval_steps_per_second": 28.429,
      "step": 12
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.8741340041160583,
      "learning_rate": 4.5e-05,
      "loss": 0.696,
      "step": 15
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.4507042253521127,
      "eval_loss": 0.700704038143158,
      "eval_runtime": 0.0371,
      "eval_samples_per_second": 1913.976,
      "eval_steps_per_second": 26.957,
      "step": 15
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.6335508227348328,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.6928,
      "step": 18
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.352112676056338,
      "eval_loss": 0.7075263857841492,
      "eval_runtime": 0.0356,
      "eval_samples_per_second": 1995.882,
      "eval_steps_per_second": 28.111,
      "step": 18
    },
    {
      "epoch": 6.0,
      "step": 18,
      "total_flos": 99911832606720.0,
      "train_loss": 0.6944943798912896,
      "train_runtime": 7.0819,
      "train_samples_per_second": 4483.276,
      "train_steps_per_second": 21.181
    }
  ],
  "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
}