checkpoint-200/trainer_state.json

{
  "best_metric": 139.37651331719127,
  "best_model_checkpoint": "./checkpoint-100",
  "epoch": 2.377581120943953,
  "global_step": 200,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "learning_rate": 3e-06,
      "loss": 4.1495,
      "step": 10
    },
    {
      "epoch": 0.24,
      "learning_rate": 6.333333333333333e-06,
      "loss": 2.9287,
      "step": 20
    },
    {
      "epoch": 0.35,
      "learning_rate": 9.666666666666667e-06,
      "loss": 2.0462,
      "step": 30
    },
    {
      "epoch": 0.47,
      "learning_rate": 9.666666666666667e-06,
      "loss": 1.6138,
      "step": 40
    },
    {
      "epoch": 0.59,
      "learning_rate": 9.296296296296296e-06,
      "loss": 1.3862,
      "step": 50
    },
    {
      "epoch": 0.71,
      "learning_rate": 8.925925925925927e-06,
      "loss": 1.2604,
      "step": 60
    },
    {
      "epoch": 0.83,
      "learning_rate": 8.555555555555556e-06,
      "loss": 1.1436,
      "step": 70
    },
    {
      "epoch": 0.94,
      "learning_rate": 8.185185185185187e-06,
      "loss": 1.168,
      "step": 80
    },
    {
      "epoch": 1.07,
      "learning_rate": 7.814814814814816e-06,
      "loss": 1.1041,
      "step": 90
    },
    {
      "epoch": 1.19,
      "learning_rate": 7.444444444444445e-06,
      "loss": 0.9683,
      "step": 100
    },
    {
      "epoch": 1.19,
      "eval_cer": 131.61659035460045,
      "eval_loss": 0.8811978697776794,
      "eval_runtime": 593.3196,
      "eval_samples_per_second": 0.863,
      "eval_steps_per_second": 0.431,
      "eval_wer": 139.37651331719127,
      "step": 100
    },
    {
      "epoch": 1.31,
      "learning_rate": 7.074074074074074e-06,
      "loss": 0.909,
      "step": 110
    },
    {
      "epoch": 1.42,
      "learning_rate": 6.703703703703704e-06,
      "loss": 0.9213,
      "step": 120
    },
    {
      "epoch": 1.54,
      "learning_rate": 6.333333333333333e-06,
      "loss": 0.9092,
      "step": 130
    },
    {
      "epoch": 1.66,
      "learning_rate": 5.962962962962963e-06,
      "loss": 0.8481,
      "step": 140
    },
    {
      "epoch": 1.78,
      "learning_rate": 5.5925925925925926e-06,
      "loss": 0.8471,
      "step": 150
    },
    {
      "epoch": 1.9,
      "learning_rate": 5.2222222222222226e-06,
      "loss": 0.8504,
      "step": 160
    },
    {
      "epoch": 2.02,
      "learning_rate": 4.851851851851852e-06,
      "loss": 0.8264,
      "step": 170
    },
    {
      "epoch": 2.14,
      "learning_rate": 4.481481481481482e-06,
      "loss": 0.7236,
      "step": 180
    },
    {
      "epoch": 2.26,
      "learning_rate": 4.111111111111111e-06,
      "loss": 0.6898,
      "step": 190
    },
    {
      "epoch": 2.38,
      "learning_rate": 3.740740740740741e-06,
      "loss": 0.6848,
      "step": 200
    },
    {
      "epoch": 2.38,
      "eval_cer": 151.33685371478225,
      "eval_loss": 0.7542899250984192,
      "eval_runtime": 551.6472,
      "eval_samples_per_second": 0.928,
      "eval_steps_per_second": 0.464,
      "eval_wer": 145.9972760290557,
      "step": 200
    }
  ],
  "max_steps": 300,
  "num_train_epochs": 4,
  "total_flos": 1.86022149046272e+18,
  "trial_name": null,
  "trial_params": null
}