last-checkpoint/trainer_state.json

{
  "best_metric": 88.17781184134891,
  "best_model_checkpoint": "./whisper-small-ha-adam-v4/checkpoint-500",
  "epoch": 3.1847133757961785,
  "eval_steps": 500,
  "global_step": 500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1592356687898089,
      "grad_norm": 21.07309341430664,
      "learning_rate": 2.5e-05,
      "loss": 4.4409,
      "step": 25
    },
    {
      "epoch": 0.3184713375796178,
      "grad_norm": 20.01163673400879,
      "learning_rate": 5e-05,
      "loss": 2.732,
      "step": 50
    },
    {
      "epoch": 0.47770700636942676,
      "grad_norm": 20.838537216186523,
      "learning_rate": 5e-05,
      "loss": 2.1765,
      "step": 75
    },
    {
      "epoch": 0.6369426751592356,
      "grad_norm": 13.77207088470459,
      "learning_rate": 5e-05,
      "loss": 1.908,
      "step": 100
    },
    {
      "epoch": 0.7961783439490446,
      "grad_norm": 13.532313346862793,
      "learning_rate": 5e-05,
      "loss": 1.745,
      "step": 125
    },
    {
      "epoch": 0.9554140127388535,
      "grad_norm": 13.651493072509766,
      "learning_rate": 5e-05,
      "loss": 1.6779,
      "step": 150
    },
    {
      "epoch": 1.1146496815286624,
      "grad_norm": 9.819719314575195,
      "learning_rate": 5e-05,
      "loss": 1.0426,
      "step": 175
    },
    {
      "epoch": 1.2738853503184713,
      "grad_norm": 9.96407699584961,
      "learning_rate": 5e-05,
      "loss": 0.7765,
      "step": 200
    },
    {
      "epoch": 1.4331210191082802,
      "grad_norm": 11.112582206726074,
      "learning_rate": 5e-05,
      "loss": 0.7978,
      "step": 225
    },
    {
      "epoch": 1.5923566878980893,
      "grad_norm": 10.045536041259766,
      "learning_rate": 5e-05,
      "loss": 0.8032,
      "step": 250
    },
    {
      "epoch": 1.7515923566878981,
      "grad_norm": 11.228167533874512,
      "learning_rate": 5e-05,
      "loss": 0.7916,
      "step": 275
    },
    {
      "epoch": 1.910828025477707,
      "grad_norm": 10.779911994934082,
      "learning_rate": 5e-05,
      "loss": 0.8241,
      "step": 300
    },
    {
      "epoch": 2.070063694267516,
      "grad_norm": 6.224853038787842,
      "learning_rate": 5e-05,
      "loss": 0.5693,
      "step": 325
    },
    {
      "epoch": 2.229299363057325,
      "grad_norm": 6.078026294708252,
      "learning_rate": 5e-05,
      "loss": 0.2365,
      "step": 350
    },
    {
      "epoch": 2.388535031847134,
      "grad_norm": 5.989273548126221,
      "learning_rate": 5e-05,
      "loss": 0.2271,
      "step": 375
    },
    {
      "epoch": 2.5477707006369426,
      "grad_norm": 8.927431106567383,
      "learning_rate": 5e-05,
      "loss": 0.2478,
      "step": 400
    },
    {
      "epoch": 2.7070063694267517,
      "grad_norm": 7.507767200469971,
      "learning_rate": 5e-05,
      "loss": 0.2846,
      "step": 425
    },
    {
      "epoch": 2.8662420382165603,
      "grad_norm": 6.9506330490112305,
      "learning_rate": 5e-05,
      "loss": 0.2722,
      "step": 450
    },
    {
      "epoch": 3.0254777070063694,
      "grad_norm": 4.77834415435791,
      "learning_rate": 5e-05,
      "loss": 0.2588,
      "step": 475
    },
    {
      "epoch": 3.1847133757961785,
      "grad_norm": 4.616095066070557,
      "learning_rate": 5e-05,
      "loss": 0.0995,
      "step": 500
    },
    {
      "epoch": 3.1847133757961785,
      "eval_loss": 1.7910184860229492,
      "eval_runtime": 264.8902,
      "eval_samples_per_second": 2.492,
      "eval_steps_per_second": 0.159,
      "eval_wer": 88.17781184134891,
      "eval_wer_ortho": 90.4296875,
      "step": 500
    }
  ],
  "logging_steps": 25,
  "max_steps": 2355,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 15,
  "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": 2.30522017775616e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}