trainer_state.json

{
  "best_global_step": 5000,
  "best_metric": 63.25362763131004,
  "best_model_checkpoint": "./whisper-tiny-ml-2/checkpoint-5000",
  "epoch": 14.0424,
  "eval_steps": 500,
  "global_step": 5000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0316,
      "grad_norm": 16.492067337036133,
      "learning_rate": 3.7275e-05,
      "loss": 1.25,
      "step": 500
    },
    {
      "epoch": 1.0316,
      "eval_loss": 0.5131270885467529,
      "eval_runtime": 464.8494,
      "eval_samples_per_second": 0.809,
      "eval_steps_per_second": 0.404,
      "eval_wer": 99.85693848354794,
      "step": 500
    },
    {
      "epoch": 2.0632,
      "grad_norm": 6.0215654373168945,
      "learning_rate": 3.335833333333333e-05,
      "loss": 0.29,
      "step": 1000
    },
    {
      "epoch": 2.0632,
      "eval_loss": 0.184208482503891,
      "eval_runtime": 453.1356,
      "eval_samples_per_second": 0.83,
      "eval_steps_per_second": 0.415,
      "eval_wer": 75.1277334968322,
      "step": 1000
    },
    {
      "epoch": 4.0264,
      "grad_norm": 4.435523986816406,
      "learning_rate": 2.9191666666666667e-05,
      "loss": 0.1325,
      "step": 1500
    },
    {
      "epoch": 4.0264,
      "eval_loss": 0.18078354001045227,
      "eval_runtime": 469.9597,
      "eval_samples_per_second": 0.8,
      "eval_steps_per_second": 0.4,
      "eval_wer": 73.65624361332516,
      "step": 1500
    },
    {
      "epoch": 5.058,
      "grad_norm": 3.5488216876983643,
      "learning_rate": 2.5025e-05,
      "loss": 0.0741,
      "step": 2000
    },
    {
      "epoch": 5.058,
      "eval_loss": 0.19534532725811005,
      "eval_runtime": 452.9853,
      "eval_samples_per_second": 0.83,
      "eval_steps_per_second": 0.415,
      "eval_wer": 71.71469446147557,
      "step": 2000
    },
    {
      "epoch": 7.0212,
      "grad_norm": 2.182008743286133,
      "learning_rate": 2.085833333333333e-05,
      "loss": 0.0424,
      "step": 2500
    },
    {
      "epoch": 7.0212,
      "eval_loss": 0.2104445993900299,
      "eval_runtime": 452.8543,
      "eval_samples_per_second": 0.83,
      "eval_steps_per_second": 0.415,
      "eval_wer": 66.13529531984467,
      "step": 2500
    },
    {
      "epoch": 8.0528,
      "grad_norm": 1.6725220680236816,
      "learning_rate": 1.6691666666666665e-05,
      "loss": 0.0241,
      "step": 3000
    },
    {
      "epoch": 8.0528,
      "eval_loss": 0.2327151745557785,
      "eval_runtime": 461.1495,
      "eval_samples_per_second": 0.815,
      "eval_steps_per_second": 0.408,
      "eval_wer": 66.17617003883099,
      "step": 3000
    },
    {
      "epoch": 10.016,
      "grad_norm": 2.495150089263916,
      "learning_rate": 1.2525e-05,
      "loss": 0.0122,
      "step": 3500
    },
    {
      "epoch": 10.016,
      "eval_loss": 0.24628408253192902,
      "eval_runtime": 457.1775,
      "eval_samples_per_second": 0.822,
      "eval_steps_per_second": 0.411,
      "eval_wer": 66.46229307173513,
      "step": 3500
    },
    {
      "epoch": 11.0476,
      "grad_norm": 0.36194851994514465,
      "learning_rate": 8.358333333333333e-06,
      "loss": 0.0054,
      "step": 4000
    },
    {
      "epoch": 11.0476,
      "eval_loss": 0.2554023861885071,
      "eval_runtime": 457.1667,
      "eval_samples_per_second": 0.822,
      "eval_steps_per_second": 0.411,
      "eval_wer": 64.41855712241978,
      "step": 4000
    },
    {
      "epoch": 13.0108,
      "grad_norm": 0.10494792461395264,
      "learning_rate": 4.191666666666667e-06,
      "loss": 0.002,
      "step": 4500
    },
    {
      "epoch": 13.0108,
      "eval_loss": 0.2660194933414459,
      "eval_runtime": 457.2487,
      "eval_samples_per_second": 0.822,
      "eval_steps_per_second": 0.411,
      "eval_wer": 64.05068465154302,
      "step": 4500
    },
    {
      "epoch": 14.0424,
      "grad_norm": 0.12081502377986908,
      "learning_rate": 2.4999999999999996e-08,
      "loss": 0.0007,
      "step": 5000
    },
    {
      "epoch": 14.0424,
      "eval_loss": 0.2695574164390564,
      "eval_runtime": 455.3759,
      "eval_samples_per_second": 0.826,
      "eval_steps_per_second": 0.413,
      "eval_wer": 63.25362763131004,
      "step": 5000
    },
    {
      "epoch": 14.0424,
      "step": 5000,
      "total_flos": 9.8544473358336e+17,
      "train_loss": 0.18335246599316596,
      "train_runtime": 6674.7567,
      "train_samples_per_second": 5.993,
      "train_steps_per_second": 0.749
    }
  ],
  "logging_steps": 500,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 9.8544473358336e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}