last-checkpoint/trainer_state.json

{
  "best_metric": 78.8027477919529,
  "best_model_checkpoint": "../KevinKibe/whisper-small-finetuned-finetuned-finetuned-finetuned-finetuned-finetuned/checkpoint-3000",
  "epoch": 1.28125,
  "eval_steps": 1000,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "grad_norm": 1.7959926128387451,
      "learning_rate": 0.0007547738693467336,
      "loss": 2.6314,
      "step": 1000
    },
    {
      "epoch": 0.25,
      "eval_loss": 2.391742706298828,
      "eval_runtime": 698.0068,
      "eval_samples_per_second": 0.358,
      "eval_steps_per_second": 0.046,
      "eval_wer": 101.71736997055937,
      "step": 1000
    },
    {
      "epoch": 1.03125,
      "grad_norm": 1.4076383113861084,
      "learning_rate": 0.0005035175879396985,
      "loss": 1.418,
      "step": 2000
    },
    {
      "epoch": 1.03125,
      "eval_loss": 2.17468523979187,
      "eval_runtime": 692.0594,
      "eval_samples_per_second": 0.361,
      "eval_steps_per_second": 0.046,
      "eval_wer": 98.62610402355251,
      "step": 2000
    },
    {
      "epoch": 1.28125,
      "grad_norm": 1.1559720039367676,
      "learning_rate": 0.00025226130653266334,
      "loss": 0.8752,
      "step": 3000
    },
    {
      "epoch": 1.28125,
      "eval_loss": 1.872373104095459,
      "eval_runtime": 475.0113,
      "eval_samples_per_second": 0.526,
      "eval_steps_per_second": 0.067,
      "eval_wer": 78.8027477919529,
      "step": 3000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 2.819342204928e+19,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}