checkpoints/checkpoint-1500/trainer_state.json

{
  "best_metric": 0.7198385688576876,
  "best_model_checkpoint": "whisper-tiny-luganda-v2/checkpoint-1000",
  "epoch": 0.3262997607135088,
  "eval_steps": 500,
  "global_step": 1500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.021753317380900587,
      "grad_norm": 47.60859680175781,
      "learning_rate": 4.5238095238095235e-06,
      "loss": 9.6365,
      "step": 100
    },
    {
      "epoch": 0.04350663476180117,
      "grad_norm": 41.293460845947266,
      "learning_rate": 9.285714285714286e-06,
      "loss": 4.282,
      "step": 200
    },
    {
      "epoch": 0.06525995214270176,
      "grad_norm": 29.167564392089844,
      "learning_rate": 1.4047619047619048e-05,
      "loss": 2.8974,
      "step": 300
    },
    {
      "epoch": 0.08701326952360235,
      "grad_norm": 30.911415100097656,
      "learning_rate": 1.880952380952381e-05,
      "loss": 2.2716,
      "step": 400
    },
    {
      "epoch": 0.10876658690450294,
      "grad_norm": 25.824464797973633,
      "learning_rate": 2.357142857142857e-05,
      "loss": 2.0498,
      "step": 500
    },
    {
      "epoch": 0.10876658690450294,
      "eval_loss": 0.9947264194488525,
      "eval_runtime": 2111.5938,
      "eval_samples_per_second": 6.458,
      "eval_steps_per_second": 1.615,
      "eval_wer": 0.8350676193361011,
      "step": 500
    },
    {
      "epoch": 0.13051990428540353,
      "grad_norm": 22.513357162475586,
      "learning_rate": 2.8333333333333335e-05,
      "loss": 1.8357,
      "step": 600
    },
    {
      "epoch": 0.1522732216663041,
      "grad_norm": 22.64017105102539,
      "learning_rate": 3.309523809523809e-05,
      "loss": 1.7019,
      "step": 700
    },
    {
      "epoch": 0.1740265390472047,
      "grad_norm": 22.659271240234375,
      "learning_rate": 3.785714285714286e-05,
      "loss": 1.6458,
      "step": 800
    },
    {
      "epoch": 0.19577985642810528,
      "grad_norm": 19.53955841064453,
      "learning_rate": 4.261904761904762e-05,
      "loss": 1.6001,
      "step": 900
    },
    {
      "epoch": 0.2175331738090059,
      "grad_norm": 21.89055061340332,
      "learning_rate": 4.738095238095238e-05,
      "loss": 1.5134,
      "step": 1000
    },
    {
      "epoch": 0.2175331738090059,
      "eval_loss": 0.7854410409927368,
      "eval_runtime": 2131.7562,
      "eval_samples_per_second": 6.397,
      "eval_steps_per_second": 1.6,
      "eval_wer": 0.7198385688576876,
      "step": 1000
    },
    {
      "epoch": 0.23928649118990647,
      "grad_norm": 24.862058639526367,
      "learning_rate": 4.988721804511278e-05,
      "loss": 1.421,
      "step": 1100
    },
    {
      "epoch": 0.26103980857080705,
      "grad_norm": 18.665178298950195,
      "learning_rate": 4.963659147869674e-05,
      "loss": 1.4254,
      "step": 1200
    },
    {
      "epoch": 0.28279312595170764,
      "grad_norm": 16.429445266723633,
      "learning_rate": 4.93859649122807e-05,
      "loss": 1.3466,
      "step": 1300
    },
    {
      "epoch": 0.3045464433326082,
      "grad_norm": 24.4464054107666,
      "learning_rate": 4.913533834586466e-05,
      "loss": 1.3342,
      "step": 1400
    },
    {
      "epoch": 0.3262997607135088,
      "grad_norm": 14.964367866516113,
      "learning_rate": 4.888471177944862e-05,
      "loss": 1.2494,
      "step": 1500
    },
    {
      "epoch": 0.3262997607135088,
      "eval_loss": 0.6914966106414795,
      "eval_runtime": 2154.6429,
      "eval_samples_per_second": 6.329,
      "eval_steps_per_second": 1.583,
      "eval_wer": 0.7817383247509934,
      "step": 1500
    }
  ],
  "logging_steps": 100,
  "max_steps": 21000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 5.9085324288e+17,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}