checkpoint-150/trainer_state.json

{
  "best_global_step": 150,
  "best_metric": 8.623494956068987,
  "best_model_checkpoint": "./whisper-finetune-small/checkpoint-150",
  "epoch": 16.685714285714287,
  "eval_steps": 50,
  "global_step": 150,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.1142857142857143,
      "grad_norm": 13.765691757202148,
      "learning_rate": 2.666666666666667e-06,
      "loss": 1.0811,
      "step": 10
    },
    {
      "epoch": 2.2285714285714286,
      "grad_norm": 8.985676765441895,
      "learning_rate": 6e-06,
      "loss": 0.5649,
      "step": 20
    },
    {
      "epoch": 3.342857142857143,
      "grad_norm": 4.251060485839844,
      "learning_rate": 9.333333333333334e-06,
      "loss": 0.255,
      "step": 30
    },
    {
      "epoch": 4.457142857142857,
      "grad_norm": 5.627696990966797,
      "learning_rate": 9.703703703703703e-06,
      "loss": 0.1037,
      "step": 40
    },
    {
      "epoch": 5.571428571428571,
      "grad_norm": 3.0673668384552,
      "learning_rate": 9.333333333333334e-06,
      "loss": 0.0595,
      "step": 50
    },
    {
      "epoch": 5.571428571428571,
      "eval_cer": 9.404490725675235,
      "eval_loss": 0.2722940146923065,
      "eval_model_preparation_time": 0.0054,
      "eval_runtime": 44.9214,
      "eval_samples_per_second": 1.358,
      "eval_steps_per_second": 0.356,
      "step": 50
    },
    {
      "epoch": 6.685714285714286,
      "grad_norm": 0.43924692273139954,
      "learning_rate": 8.962962962962963e-06,
      "loss": 0.0276,
      "step": 60
    },
    {
      "epoch": 7.8,
      "grad_norm": 0.8204056024551392,
      "learning_rate": 8.592592592592593e-06,
      "loss": 0.0137,
      "step": 70
    },
    {
      "epoch": 8.914285714285715,
      "grad_norm": 0.2582692801952362,
      "learning_rate": 8.222222222222222e-06,
      "loss": 0.0099,
      "step": 80
    },
    {
      "epoch": 10.0,
      "grad_norm": 1.72308349609375,
      "learning_rate": 7.851851851851853e-06,
      "loss": 0.0068,
      "step": 90
    },
    {
      "epoch": 11.114285714285714,
      "grad_norm": 0.21061824262142181,
      "learning_rate": 7.481481481481482e-06,
      "loss": 0.005,
      "step": 100
    },
    {
      "epoch": 11.114285714285714,
      "eval_cer": 8.688577936869509,
      "eval_loss": 0.27061235904693604,
      "eval_model_preparation_time": 0.0054,
      "eval_runtime": 41.153,
      "eval_samples_per_second": 1.482,
      "eval_steps_per_second": 0.389,
      "step": 100
    },
    {
      "epoch": 12.228571428571428,
      "grad_norm": 0.053500544279813766,
      "learning_rate": 7.111111111111112e-06,
      "loss": 0.0024,
      "step": 110
    },
    {
      "epoch": 13.342857142857143,
      "grad_norm": 0.01863274723291397,
      "learning_rate": 6.740740740740741e-06,
      "loss": 0.0021,
      "step": 120
    },
    {
      "epoch": 14.457142857142857,
      "grad_norm": 0.12402216345071793,
      "learning_rate": 6.370370370370371e-06,
      "loss": 0.0016,
      "step": 130
    },
    {
      "epoch": 15.571428571428571,
      "grad_norm": 0.08812814950942993,
      "learning_rate": 6e-06,
      "loss": 0.0011,
      "step": 140
    },
    {
      "epoch": 16.685714285714287,
      "grad_norm": 0.01754395291209221,
      "learning_rate": 5.62962962962963e-06,
      "loss": 0.0009,
      "step": 150
    },
    {
      "epoch": 16.685714285714287,
      "eval_cer": 8.623494956068987,
      "eval_loss": 0.28332972526550293,
      "eval_model_preparation_time": 0.0054,
      "eval_runtime": 40.2364,
      "eval_samples_per_second": 1.516,
      "eval_steps_per_second": 0.398,
      "step": 150
    }
  ],
  "logging_steps": 10,
  "max_steps": 300,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 38,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 6.7413549514752e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}