trainer_state.json

{
  "best_metric": 26.972262247838618,
  "best_model_checkpoint": "./whisper-medium-hy-large-data\\checkpoint-3090",
  "epoch": 8.0,
  "eval_steps": 500,
  "global_step": 3090,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.9993527508090615,
      "grad_norm": 2.3409805297851562,
      "learning_rate": 9.60099492175355e-06,
      "loss": 0.0122,
      "step": 386
    },
    {
      "epoch": 0.9993527508090615,
      "eval_loss": 0.19175899028778076,
      "eval_runtime": 42548.1103,
      "eval_samples_per_second": 0.101,
      "eval_steps_per_second": 0.05,
      "eval_wer": 28.368155619596543,
      "step": 386
    },
    {
      "epoch": 1.998705501618123,
      "grad_norm": 1.1300331354141235,
      "learning_rate": 9.200953466680485e-06,
      "loss": 0.0095,
      "step": 772
    },
    {
      "epoch": 1.998705501618123,
      "eval_loss": 0.19632793962955475,
      "eval_runtime": 45633.7169,
      "eval_samples_per_second": 0.094,
      "eval_steps_per_second": 0.047,
      "eval_wer": 30.382444764649374,
      "step": 772
    },
    {
      "epoch": 2.9980582524271844,
      "grad_norm": 0.9480321407318115,
      "learning_rate": 8.800912011607421e-06,
      "loss": 0.0076,
      "step": 1158
    },
    {
      "epoch": 2.9980582524271844,
      "eval_loss": 0.19936133921146393,
      "eval_runtime": 45880.7942,
      "eval_samples_per_second": 0.093,
      "eval_steps_per_second": 0.047,
      "eval_wer": 27.785782901056677,
      "step": 1158
    },
    {
      "epoch": 4.0,
      "grad_norm": 4.70139217376709,
      "learning_rate": 8.399834179707742e-06,
      "loss": 0.0057,
      "step": 1545
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.20888063311576843,
      "eval_runtime": 32746.3235,
      "eval_samples_per_second": 0.131,
      "eval_steps_per_second": 0.065,
      "eval_wer": 27.668707973102784,
      "step": 1545
    },
    {
      "epoch": 4.999352750809061,
      "grad_norm": 1.9372472763061523,
      "learning_rate": 7.999792724634678e-06,
      "loss": 0.0046,
      "step": 1931
    },
    {
      "epoch": 4.999352750809061,
      "eval_loss": 0.2178862988948822,
      "eval_runtime": 46002.5276,
      "eval_samples_per_second": 0.093,
      "eval_steps_per_second": 0.047,
      "eval_wer": 28.542267050912585,
      "step": 1931
    },
    {
      "epoch": 5.998705501618123,
      "grad_norm": 1.5416231155395508,
      "learning_rate": 7.599751269561612e-06,
      "loss": 0.0036,
      "step": 2317
    },
    {
      "epoch": 5.998705501618123,
      "eval_loss": 0.21689267456531525,
      "eval_runtime": 45863.2972,
      "eval_samples_per_second": 0.093,
      "eval_steps_per_second": 0.047,
      "eval_wer": 27.554634966378483,
      "step": 2317
    },
    {
      "epoch": 6.998058252427184,
      "grad_norm": 0.7119281888008118,
      "learning_rate": 7.199709814488549e-06,
      "loss": 0.0028,
      "step": 2703
    },
    {
      "epoch": 6.998058252427184,
      "eval_loss": 0.21707293391227722,
      "eval_runtime": 45784.2835,
      "eval_samples_per_second": 0.094,
      "eval_steps_per_second": 0.047,
      "eval_wer": 27.251440922190202,
      "step": 2703
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.8986366391181946,
      "learning_rate": 6.798631982588871e-06,
      "loss": 0.0028,
      "step": 3090
    },
    {
      "epoch": 8.0,
      "eval_loss": 0.2156277447938919,
      "eval_runtime": 32539.3621,
      "eval_samples_per_second": 0.132,
      "eval_steps_per_second": 0.066,
      "eval_wer": 26.972262247838618,
      "step": 3090
    }
  ],
  "logging_steps": 500,
  "max_steps": 9650,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 25,
  "save_steps": 500,
  "total_flos": 5.04587066277888e+19,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}