trainer_state.json

{
  "best_metric": 0.6951221227645874,
  "best_model_checkpoint": "tiny_bert_km_50_v1_rte/checkpoint-10",
  "epoch": 6.0,
  "eval_steps": 500,
  "global_step": 60,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.6421326398849487,
      "learning_rate": 4.9e-05,
      "loss": 0.6977,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5054151624548736,
      "eval_loss": 0.6951221227645874,
      "eval_runtime": 0.11,
      "eval_samples_per_second": 2518.222,
      "eval_steps_per_second": 18.182,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.6818793416023254,
      "learning_rate": 4.8e-05,
      "loss": 0.684,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.49458483754512633,
      "eval_loss": 0.7016090154647827,
      "eval_runtime": 0.1147,
      "eval_samples_per_second": 2414.683,
      "eval_steps_per_second": 17.435,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6151508092880249,
      "learning_rate": 4.7e-05,
      "loss": 0.6682,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5018050541516246,
      "eval_loss": 0.7009462714195251,
      "eval_runtime": 0.1166,
      "eval_samples_per_second": 2376.571,
      "eval_steps_per_second": 17.159,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.5758317708969116,
      "learning_rate": 4.600000000000001e-05,
      "loss": 0.6498,
      "step": 40
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.48736462093862815,
      "eval_loss": 0.7171741127967834,
      "eval_runtime": 0.143,
      "eval_samples_per_second": 1937.287,
      "eval_steps_per_second": 13.988,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.9044778943061829,
      "learning_rate": 4.5e-05,
      "loss": 0.6256,
      "step": 50
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.5126353790613718,
      "eval_loss": 0.7283922433853149,
      "eval_runtime": 0.1117,
      "eval_samples_per_second": 2479.088,
      "eval_steps_per_second": 17.9,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.8710564374923706,
      "learning_rate": 4.4000000000000006e-05,
      "loss": 0.5891,
      "step": 60
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.5054151624548736,
      "eval_loss": 0.7591274380683899,
      "eval_runtime": 0.1101,
      "eval_samples_per_second": 2516.177,
      "eval_steps_per_second": 18.167,
      "step": 60
    },
    {
      "epoch": 6.0,
      "step": 60,
      "total_flos": 391780256993280.0,
      "train_loss": 0.6523992617925009,
      "train_runtime": 15.3656,
      "train_samples_per_second": 8102.502,
      "train_steps_per_second": 32.54
    }
  ],
  "logging_steps": 1,
  "max_steps": 500,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 50,
  "save_steps": 500,
  "total_flos": 391780256993280.0,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}