trainer_state.json

{
  "best_metric": 35.37824487651077,
  "best_model_checkpoint": "/models/hfhub/whisper-tiny-ft-cy/checkpoint-4000",
  "epoch": 1.005593614480548,
  "eval_steps": 1000,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06,
      "learning_rate": 5e-06,
      "loss": 1.8194,
      "step": 250
    },
    {
      "epoch": 0.13,
      "learning_rate": 1e-05,
      "loss": 0.9849,
      "step": 500
    },
    {
      "epoch": 0.19,
      "learning_rate": 9.285714285714288e-06,
      "loss": 0.7915,
      "step": 750
    },
    {
      "epoch": 0.25,
      "learning_rate": 8.571428571428571e-06,
      "loss": 0.7059,
      "step": 1000
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.73893803358078,
      "eval_runtime": 4402.4141,
      "eval_samples_per_second": 5.958,
      "eval_steps_per_second": 0.745,
      "eval_wer": 44.47125591171834,
      "step": 1000
    },
    {
      "epoch": 0.31,
      "learning_rate": 7.857142857142858e-06,
      "loss": 0.6376,
      "step": 1250
    },
    {
      "epoch": 0.38,
      "learning_rate": 7.1428571428571436e-06,
      "loss": 0.6106,
      "step": 1500
    },
    {
      "epoch": 0.44,
      "learning_rate": 6.4285714285714295e-06,
      "loss": 0.5788,
      "step": 1750
    },
    {
      "epoch": 0.5,
      "learning_rate": 5.7142857142857145e-06,
      "loss": 0.5534,
      "step": 2000
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.6115749478340149,
      "eval_runtime": 4567.6465,
      "eval_samples_per_second": 5.743,
      "eval_steps_per_second": 0.718,
      "eval_wer": 39.26852338413032,
      "step": 2000
    },
    {
      "epoch": 0.57,
      "learning_rate": 5e-06,
      "loss": 0.5487,
      "step": 2250
    },
    {
      "epoch": 0.63,
      "learning_rate": 4.2857142857142855e-06,
      "loss": 0.5209,
      "step": 2500
    },
    {
      "epoch": 0.69,
      "learning_rate": 3.5714285714285718e-06,
      "loss": 0.5085,
      "step": 2750
    },
    {
      "epoch": 0.75,
      "learning_rate": 2.8571428571428573e-06,
      "loss": 0.5087,
      "step": 3000
    },
    {
      "epoch": 0.75,
      "eval_loss": 0.5615290403366089,
      "eval_runtime": 4330.9978,
      "eval_samples_per_second": 6.057,
      "eval_steps_per_second": 0.757,
      "eval_wer": 36.01555438780872,
      "step": 3000
    },
    {
      "epoch": 0.82,
      "learning_rate": 2.1428571428571427e-06,
      "loss": 0.493,
      "step": 3250
    },
    {
      "epoch": 0.88,
      "learning_rate": 1.4285714285714286e-06,
      "loss": 0.4827,
      "step": 3500
    },
    {
      "epoch": 0.94,
      "learning_rate": 7.142857142857143e-07,
      "loss": 0.4842,
      "step": 3750
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0,
      "loss": 0.4749,
      "step": 4000
    },
    {
      "epoch": 1.01,
      "eval_loss": 0.5461679100990295,
      "eval_runtime": 4286.2436,
      "eval_samples_per_second": 6.12,
      "eval_steps_per_second": 0.765,
      "eval_wer": 35.37824487651077,
      "step": 4000
    },
    {
      "epoch": 1.01,
      "step": 4000,
      "total_flos": 3.15119267647488e+18,
      "train_loss": 0.6689802742004395,
      "train_runtime": 33479.3884,
      "train_samples_per_second": 3.823,
      "train_steps_per_second": 0.119
    },
    {
      "epoch": 1.01,
      "eval_loss": 0.5461679100990295,
      "eval_runtime": 4251.2772,
      "eval_samples_per_second": 6.17,
      "eval_steps_per_second": 0.771,
      "eval_wer": 35.37824487651077,
      "step": 4000
    }
  ],
  "logging_steps": 250,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 1000,
  "total_flos": 3.15119267647488e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}