File size: 19,056 Bytes
29cdc30 96f1a15 29cdc30 96f1a15 29cdc30 96f1a15 29cdc30 96f1a15 29cdc30 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 |
---
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: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# 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: 0.4208
- Wer: 18.5421
- Avg Precision Exact: 0.8482
- Avg Recall Exact: 0.8623
- Avg F1 Exact: 0.8545
- Avg Precision Letter Shift: 0.8778
- Avg Recall Letter Shift: 0.8926
- Avg F1 Letter Shift: 0.8844
- Avg Precision Word Level: 0.8803
- Avg Recall Word Level: 0.8946
- Avg F1 Word Level: 0.8866
- Avg Precision Word Shift: 0.9416
- Avg Recall Word Shift: 0.9549
- Avg F1 Word Shift: 0.9475
- Precision Median Exact: 0.9167
- Recall Median Exact: 0.9167
- F1 Median Exact: 0.9167
- 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: 0.6154
- Recall Min Word Shift: 0.6667
- F1 Min Word Shift: 0.64
## 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: 500
- training_steps: 200000
- 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 |
|:-------------:|:-----:|:------:|:---------------:|:--------:|:-------------------:|:----------------:|:------------:|:--------------------------:|:-----------------------:|:-------------------:|:------------------------:|:---------------------:|:-----------------:|:------------------------:|:---------------------:|:-----------------:|:----------------------:|:-------------------:|:---------------:|:-------------------:|:----------------:|:------------:|:-------------------:|:----------------:|:------------:|:--------------------------:|:-----------------------:|:-------------------:|:------------------------:|:---------------------:|:-----------------:|:------------------------:|:---------------------:|:-----------------:|
| No log | 8e-05 | 1 | 7.8969 | 120.8943 | 0.0002 | 0.0000 | 0.0000 | 0.0030 | 0.0028 | 0.0027 | 0.0024 | 0.0137 | 0.0027 | 0.0141 | 0.0233 | 0.0133 | 0.0 | 0.0 | 0.0 | 0.0625 | 0.0074 | 0.0132 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 0.0777 | 0.8 | 10000 | 0.2194 | 22.7167 | 0.7902 | 0.7985 | 0.7935 | 0.8205 | 0.8298 | 0.8243 | 0.8264 | 0.8358 | 0.8303 | 0.9175 | 0.9318 | 0.9237 | 0.9091 | 0.9091 | 0.8889 | 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 |
| 0.0363 | 1.6 | 20000 | 0.2639 | 19.9619 | 0.8266 | 0.8330 | 0.8289 | 0.8554 | 0.8623 | 0.8580 | 0.8620 | 0.8676 | 0.8640 | 0.9419 | 0.9519 | 0.9459 | 0.9091 | 0.9167 | 0.9091 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.6154 | 0.6429 | 0.64 |
| 0.0189 | 2.4 | 30000 | 0.2936 | 20.0678 | 0.8200 | 0.8284 | 0.8233 | 0.8485 | 0.8574 | 0.8521 | 0.8538 | 0.8621 | 0.8571 | 0.9365 | 0.9491 | 0.9419 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.4167 | 0.4545 | 0.4348 |
| 0.0089 | 3.2 | 40000 | 0.3115 | 19.8347 | 0.8154 | 0.8264 | 0.8200 | 0.8456 | 0.8570 | 0.8504 | 0.8488 | 0.8601 | 0.8536 | 0.9396 | 0.9526 | 0.9452 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.5 | 0.5556 | 0.5263 |
| 0.0095 | 4.0 | 50000 | 0.3330 | 20.0042 | 0.8325 | 0.8425 | 0.8366 | 0.8628 | 0.8739 | 0.8674 | 0.8668 | 0.8780 | 0.8715 | 0.9376 | 0.9481 | 0.9418 | 0.9091 | 0.9167 | 0.9091 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.5385 | 0.5833 | 0.5600 |
| 0.0062 | 4.8 | 60000 | 0.3605 | 19.6228 | 0.8291 | 0.8399 | 0.8336 | 0.8619 | 0.8733 | 0.8667 | 0.8651 | 0.8752 | 0.8693 | 0.9376 | 0.9466 | 0.9413 | 0.9091 | 0.9167 | 0.9167 | 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 | 0.0833 | 0.0952 |
| 0.0052 | 5.6 | 70000 | 0.3855 | 20.4492 | 0.8137 | 0.8244 | 0.8182 | 0.8398 | 0.8513 | 0.8448 | 0.8446 | 0.8568 | 0.8499 | 0.9330 | 0.9448 | 0.9380 | 0.9091 | 0.9091 | 0.9 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.4545 | 0.5556 | 0.5 |
| 0.0018 | 6.4 | 80000 | 0.3755 | 20.1950 | 0.8275 | 0.8378 | 0.8318 | 0.8579 | 0.8695 | 0.8628 | 0.8617 | 0.8732 | 0.8666 | 0.9402 | 0.9511 | 0.9447 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.6667 | 0.6 | 0.6316 |
| 0.0038 | 7.2 | 90000 | 0.3829 | 20.4281 | 0.8308 | 0.8429 | 0.8359 | 0.8586 | 0.8715 | 0.8640 | 0.8619 | 0.8738 | 0.8669 | 0.9328 | 0.9467 | 0.9387 | 0.9167 | 0.9167 | 0.9167 | 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 | 0.0769 | 0.08 |
| 0.0029 | 8.0 | 100000 | 0.3823 | 19.4533 | 0.8423 | 0.8504 | 0.8455 | 0.8686 | 0.8773 | 0.8720 | 0.8716 | 0.8795 | 0.8747 | 0.9435 | 0.9498 | 0.9457 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.6154 | 0.6667 | 0.64 |
| 0.0013 | 8.8 | 110000 | 0.3992 | 19.7499 | 0.8323 | 0.8449 | 0.8378 | 0.8609 | 0.8745 | 0.8668 | 0.8639 | 0.8767 | 0.8694 | 0.9404 | 0.9532 | 0.9459 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.6154 | 0.625 | 0.625 |
| 0.0015 | 9.6 | 120000 | 0.3963 | 19.8135 | 0.8317 | 0.8426 | 0.8364 | 0.8629 | 0.8745 | 0.8679 | 0.8660 | 0.8775 | 0.8710 | 0.9355 | 0.9460 | 0.9399 | 0.9167 | 0.9167 | 0.9167 | 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 |
| 0.0005 | 10.4 | 130000 | 0.4010 | 19.3261 | 0.8398 | 0.8500 | 0.8441 | 0.8705 | 0.8816 | 0.8751 | 0.8737 | 0.8845 | 0.8782 | 0.9379 | 0.9497 | 0.9430 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.2727 | 0.2727 | 0.2727 |
| 0.0003 | 11.2 | 140000 | 0.4015 | 19.3685 | 0.8386 | 0.8518 | 0.8444 | 0.8676 | 0.8817 | 0.8738 | 0.8714 | 0.8844 | 0.8771 | 0.9319 | 0.9461 | 0.9381 | 0.9091 | 0.9167 | 0.9091 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.2727 | 0.2727 | 0.2727 |
| 0.0007 | 12.0 | 150000 | 0.4071 | 19.2837 | 0.8341 | 0.8455 | 0.8389 | 0.8642 | 0.8765 | 0.8695 | 0.8684 | 0.8802 | 0.8735 | 0.9380 | 0.9493 | 0.9428 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.5833 | 0.5385 | 0.5600 |
| 0.0008 | 12.8 | 160000 | 0.4111 | 19.3685 | 0.8391 | 0.8491 | 0.8433 | 0.8683 | 0.8791 | 0.8729 | 0.8722 | 0.8824 | 0.8765 | 0.9373 | 0.9494 | 0.9425 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.5833 | 0.5833 | 0.5833 |
| 0.0002 | 13.6 | 170000 | 0.4107 | 19.4745 | 0.8295 | 0.8400 | 0.8339 | 0.8607 | 0.8721 | 0.8655 | 0.8645 | 0.8751 | 0.8689 | 0.9379 | 0.9493 | 0.9427 | 0.9091 | 0.9167 | 0.9091 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.5833 | 0.5833 | 0.5833 |
| 0.0001 | 14.4 | 180000 | 0.4214 | 18.7752 | 0.8473 | 0.8609 | 0.8533 | 0.8780 | 0.8925 | 0.8844 | 0.8810 | 0.8940 | 0.8867 | 0.9407 | 0.9539 | 0.9464 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.6364 | 0.7273 | 0.6957 |
| 0.0001 | 15.2 | 190000 | 0.4146 | 18.7116 | 0.8474 | 0.8609 | 0.8533 | 0.8768 | 0.8910 | 0.8831 | 0.8804 | 0.8944 | 0.8866 | 0.9413 | 0.9550 | 0.9473 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.6667 | 0.7273 | 0.6957 |
| 0.0005 | 16.0 | 200000 | 0.4208 | 18.5421 | 0.8482 | 0.8623 | 0.8545 | 0.8778 | 0.8926 | 0.8844 | 0.8803 | 0.8946 | 0.8866 | 0.9416 | 0.9549 | 0.9475 | 0.9167 | 0.9167 | 0.9167 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.6154 | 0.6667 | 0.64 |
### Framework versions
- Transformers 4.41.2
- Pytorch 2.2.1
- Datasets 2.20.0
- Tokenizers 0.19.1
|