--- language: - he license: apache-2.0 base_model: openai/whisper-small tags: - hf-asr-leaderboard - generated_from_trainer metrics: - wer model-index: - name: he-cantillation results: [] --- # he-cantillation This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 0.1579 - Wer: 12.5795 - Avg Precision Exact: 0.8918 - Avg Recall Exact: 0.8920 - Avg F1 Exact: 0.8915 - Avg Precision Letter Shift: 0.9121 - Avg Recall Letter Shift: 0.9126 - Avg F1 Letter Shift: 0.9120 - Avg Precision Word Level: 0.9146 - Avg Recall Word Level: 0.9150 - Avg F1 Word Level: 0.9144 - Avg Precision Word Shift: 0.9737 - Avg Recall Word Shift: 0.9750 - Avg F1 Word Shift: 0.9739 - Precision Median Exact: 1.0 - Recall Median Exact: 1.0 - F1 Median Exact: 1.0 - Precision Max Exact: 1.0 - Recall Max Exact: 1.0 - F1 Max Exact: 1.0 - Precision Min Exact: 0.0 - Recall Min Exact: 0.0 - F1 Min Exact: 0.0 - Precision Min Letter Shift: 0.0 - Recall Min Letter Shift: 0.0 - F1 Min Letter Shift: 0.0 - Precision Min Word Level: 0.0 - Recall Min Word Level: 0.0 - F1 Min Word Level: 0.0 - Precision Min Word Shift: 0.1429 - Recall Min Word Shift: 0.1 - F1 Min Word Shift: 0.1176 ## 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: - learning_rate: 1e-05 - train_batch_size: 8 - eval_batch_size: 32 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - lr_scheduler_warmup_steps: 1000 - training_steps: 50000 - mixed_precision_training: Native AMP ### Training results | Training Loss | Epoch | Step | Validation Loss | Wer | Avg Precision Exact | Avg Recall Exact | Avg F1 Exact | Avg Precision Letter Shift | Avg Recall Letter Shift | Avg F1 Letter Shift | Avg Precision Word Level | Avg Recall Word Level | Avg F1 Word Level | Avg Precision Word Shift | Avg Recall Word Shift | Avg F1 Word Shift | Precision Median Exact | Recall Median Exact | F1 Median Exact | Precision Max Exact | Recall Max Exact | F1 Max Exact | Precision Min Exact | Recall Min Exact | F1 Min Exact | Precision Min Letter Shift | Recall Min Letter Shift | F1 Min Letter Shift | Precision Min Word Level | Recall Min Word Level | F1 Min Word Level | Precision Min Word Shift | Recall Min Word Shift | F1 Min Word Shift | |:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------------------:|:----------------:|:------------:|:--------------------------:|:-----------------------:|:-------------------:|:------------------------:|:---------------------:|:-----------------:|:------------------------:|:---------------------:|:-----------------:|:----------------------:|:-------------------:|:---------------:|:-------------------:|:----------------:|:------------:|:-------------------:|:----------------:|:------------:|:--------------------------:|:-----------------------:|:-------------------:|:------------------------:|:---------------------:|:-----------------:|:------------------------:|:---------------------:|:-----------------:| | 0.0443 | 0.4 | 5000 | 0.1428 | 18.7251 | 0.8438 | 0.8491 | 0.8459 | 0.8710 | 0.8764 | 0.8731 | 0.8744 | 0.8797 | 0.8765 | 0.9469 | 0.9529 | 0.9492 | 0.9167 | 0.9231 | 0.9231 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | | 0.0165 | 0.8 | 10000 | 0.1359 | 15.5691 | 0.8644 | 0.8668 | 0.8652 | 0.8882 | 0.8907 | 0.8890 | 0.8912 | 0.8938 | 0.8920 | 0.9621 | 0.9653 | 0.9632 | 0.9310 | 0.9375 | 0.9474 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1429 | 0.1 | 0.1176 | | 0.0282 | 1.2 | 15000 | 0.1395 | 14.6859 | 0.8710 | 0.8721 | 0.8712 | 0.8930 | 0.8941 | 0.8931 | 0.8960 | 0.8971 | 0.8961 | 0.9660 | 0.9672 | 0.9662 | 0.9375 | 0.9412 | 0.9524 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1111 | 0.1 | 0.1176 | | 0.0076 | 1.6 | 20000 | 0.1468 | 13.9542 | 0.8822 | 0.8850 | 0.8832 | 0.9040 | 0.9069 | 0.9051 | 0.9067 | 0.9098 | 0.9078 | 0.9672 | 0.9717 | 0.9689 | 0.9375 | 1.0 | 0.9565 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1429 | 0.1 | 0.1176 | | 0.0075 | 2.0 | 25000 | 0.1491 | 13.6179 | 0.8854 | 0.8860 | 0.8853 | 0.9067 | 0.9073 | 0.9066 | 0.9092 | 0.9101 | 0.9092 | 0.9685 | 0.9698 | 0.9686 | 1.0 | 1.0 | 0.9630 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1429 | 0.1 | 0.1176 | | 0.0033 | 2.4 | 30000 | 0.1531 | 13.2114 | 0.8901 | 0.8900 | 0.8897 | 0.9102 | 0.9101 | 0.9097 | 0.9126 | 0.9127 | 0.9122 | 0.9712 | 0.9723 | 0.9713 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1429 | 0.1 | 0.1176 | | 0.0045 | 2.8 | 35000 | 0.1535 | 13.0488 | 0.8920 | 0.8921 | 0.8916 | 0.9129 | 0.9131 | 0.9126 | 0.9153 | 0.9155 | 0.9150 | 0.9726 | 0.9734 | 0.9725 | 1.0 | 1.0 | 0.9697 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1429 | 0.1 | 0.1176 | | 0.0011 | 3.2 | 40000 | 0.1543 | 12.9084 | 0.8924 | 0.8940 | 0.8928 | 0.9131 | 0.9149 | 0.9136 | 0.9158 | 0.9174 | 0.9162 | 0.9713 | 0.9735 | 0.9719 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0833 | 0.0909 | 0.0870 | | 0.0007 | 3.6 | 45000 | 0.1571 | 12.7679 | 0.8907 | 0.8926 | 0.8913 | 0.9118 | 0.9140 | 0.9125 | 0.9146 | 0.9165 | 0.9152 | 0.9724 | 0.9750 | 0.9733 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1429 | 0.1 | 0.1176 | | 0.0005 | 4.0 | 50000 | 0.1579 | 12.5795 | 0.8918 | 0.8920 | 0.8915 | 0.9121 | 0.9126 | 0.9120 | 0.9146 | 0.9150 | 0.9144 | 0.9737 | 0.9750 | 0.9739 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1429 | 0.1 | 0.1176 | ### Framework versions - Transformers 4.39.0.dev0 - Pytorch 2.2.1+cu121 - Datasets 2.16.1 - Tokenizers 0.15.0