whisper3_0020 / README.md
bigmorning's picture
Upload TFWhisperForConditionalGeneration
7af2488
metadata
license: apache-2.0
tags:
  - generated_from_keras_callback
model-index:
  - name: whisper3_0020
    results: []

whisper3_0020

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: 0.1844
  • Train Accuracy: 0.0334
  • Validation Loss: 0.5619
  • Validation Accuracy: 0.0313
  • Epoch: 19

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 Validation Loss Validation Accuracy Epoch
5.0832 0.0116 4.4298 0.0124 0
4.3130 0.0131 4.0733 0.0141 1
3.9211 0.0146 3.6762 0.0157 2
3.5505 0.0159 3.3453 0.0171 3
3.1592 0.0175 2.8062 0.0199 4
2.2581 0.0220 1.7622 0.0252 5
1.4671 0.0259 1.2711 0.0276 6
1.0779 0.0278 1.0220 0.0288 7
0.8591 0.0290 0.8836 0.0295 8
0.7159 0.0297 0.7918 0.0300 9
0.6105 0.0304 0.7276 0.0303 10
0.5287 0.0309 0.6850 0.0306 11
0.4614 0.0313 0.6472 0.0308 12
0.4049 0.0317 0.6199 0.0310 13
0.3562 0.0320 0.6019 0.0311 14
0.3139 0.0324 0.5868 0.0311 15
0.2766 0.0326 0.5751 0.0312 16
0.2438 0.0329 0.5701 0.0312 17
0.2116 0.0332 0.5686 0.0313 18
0.1844 0.0334 0.5619 0.0313 19

Framework versions

  • Transformers 4.25.0.dev0
  • TensorFlow 2.9.2
  • Datasets 2.6.1
  • Tokenizers 0.13.2