bigmorning's picture
Upload TFWhisperForConditionalGeneration
0610fb1
metadata
license: apache-2.0
base_model: openai/whisper-tiny
tags:
  - generated_from_keras_callback
model-index:
  - name: whisper_charsplit_0005
    results: []

whisper_charsplit_0005

This model is a fine-tuned version of openai/whisper-tiny on an unknown dataset. It achieves the following results on the evaluation set:

  • Train Loss: 1.0544
  • Train Accuracy: 0.0561
  • Train Wermet: 9.9365
  • Validation Loss: 0.8809
  • Validation Accuracy: 0.0623
  • Validation Wermet: 10.1087
  • Epoch: 4

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • optimizer: {'name': 'AdamWeightDecay', 'learning_rate': 1e-05, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False, 'weight_decay_rate': 0.01}
  • training_precision: float32

Training results

Train Loss Train Accuracy Train Wermet Validation Loss Validation Accuracy Validation Wermet Epoch
2.7836 0.0245 8.8338 2.1866 0.0348 6.5008 0
2.0715 0.0354 7.5148 1.8725 0.0410 5.8800 1
1.7730 0.0412 7.4995 1.5964 0.0467 6.7257 2
1.4468 0.0478 8.1713 1.2401 0.0544 8.7249 3
1.0544 0.0561 9.9365 0.8809 0.0623 10.1087 4

Framework versions

  • Transformers 4.32.0.dev0
  • TensorFlow 2.12.0
  • Tokenizers 0.13.3