whisper_0015 / README.md
bigmorning's picture
Upload TFWhisperForConditionalGeneration
62bd1e2
metadata
license: apache-2.0
tags:
  - generated_from_keras_callback
model-index:
  - name: whisper_0015
    results: []

whisper_0015

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: 0.3281
  • Train Accuracy: 0.0322
  • Validation Loss: 0.5841
  • Validation Accuracy: 0.0311
  • Epoch: 14

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
5.0856 0.0116 4.4440 0.0123 0
4.3149 0.0131 4.0521 0.0142 1
3.9260 0.0146 3.7264 0.0153 2
3.5418 0.0160 3.3026 0.0174 3
2.7510 0.0198 2.0157 0.0241 4
1.6782 0.0250 1.3567 0.0273 5
1.1705 0.0274 1.0678 0.0286 6
0.9126 0.0287 0.9152 0.0294 7
0.7514 0.0296 0.8057 0.0299 8
0.6371 0.0302 0.7409 0.0302 9
0.5498 0.0307 0.6854 0.0306 10
0.4804 0.0312 0.6518 0.0307 11
0.4214 0.0316 0.6200 0.0310 12
0.3713 0.0319 0.5947 0.0311 13
0.3281 0.0322 0.5841 0.0311 14

Framework versions

  • Transformers 4.25.0.dev0
  • TensorFlow 2.9.2
  • Tokenizers 0.13.2