File size: 19,198 Bytes
37757db |
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 |
---
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: Teamim-small_WeightDecay-0.05_Augmented_Old-Data_date-21-07-2024_14-34_WithNikud
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. -->
# Teamim-small_WeightDecay-0.05_Augmented_Old-Data_date-21-07-2024_14-34_WithNikud
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: 0.2584
- Wer: 17.4613
- Avg Precision Exact: 0.8722
- Avg Recall Exact: 0.8838
- Avg F1 Exact: 0.8772
- Avg Precision Letter Shift: 0.8981
- Avg Recall Letter Shift: 0.9102
- Avg F1 Letter Shift: 0.9033
- Avg Precision Word Level: 0.9008
- Avg Recall Word Level: 0.9121
- Avg F1 Word Level: 0.9057
- Avg Precision Word Shift: 0.9601
- Avg Recall Word Shift: 0.9730
- Avg F1 Word Shift: 0.9657
- Precision Median Exact: 0.9286
- Recall Median Exact: 0.9333
- F1 Median Exact: 0.9565
- 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: 0.6364
- Recall Min Word Shift: 0.6364
- F1 Min Word Shift: 0.6364
## 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 | 0.0001 | 1 | 10.1840 | 103.2210 | 0.0002 | 0.0000 | 0.0000 | 0.0022 | 0.0069 | 0.0014 | 0.0019 | 0.0169 | 0.0033 | 0.0094 | 0.0316 | 0.0107 | 0.0 | 0.0 | 0.0 | 0.0714 | 0.0069 | 0.0127 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 |
| 0.0146 | 1.1457 | 10000 | 0.1688 | 23.0557 | 0.8202 | 0.8417 | 0.8300 | 0.8524 | 0.8746 | 0.8625 | 0.8584 | 0.8783 | 0.8674 | 0.9348 | 0.9545 | 0.9436 | 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 | 0.2308 | 0.375 | 0.2857 |
| 0.0105 | 2.2915 | 20000 | 0.1836 | 20.5764 | 0.8526 | 0.8616 | 0.8563 | 0.8800 | 0.8897 | 0.8840 | 0.8816 | 0.8917 | 0.8858 | 0.9510 | 0.9624 | 0.9559 | 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 | 0.5556 | 0.625 | 0.5882 |
| 0.0051 | 3.4372 | 30000 | 0.1949 | 19.7288 | 0.8570 | 0.8663 | 0.8608 | 0.8812 | 0.8912 | 0.8854 | 0.8834 | 0.8928 | 0.8872 | 0.9540 | 0.9642 | 0.9583 | 0.9231 | 0.9286 | 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 | 0.6429 | 0.6923 | 0.6667 |
| 0.0025 | 4.5830 | 40000 | 0.2128 | 19.5380 | 0.8631 | 0.8765 | 0.8690 | 0.8908 | 0.9051 | 0.8971 | 0.8951 | 0.9065 | 0.9000 | 0.9514 | 0.9654 | 0.9577 | 0.9231 | 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 | 0.6364 | 0.6923 | 0.6667 |
| 0.003 | 5.7287 | 50000 | 0.2182 | 20.0042 | 0.8632 | 0.8761 | 0.8689 | 0.8901 | 0.9039 | 0.8962 | 0.8935 | 0.9057 | 0.8988 | 0.9539 | 0.9671 | 0.9597 | 0.9231 | 0.9231 | 0.9286 | 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 |
| 0.0027 | 6.8744 | 60000 | 0.2184 | 19.0083 | 0.8532 | 0.8641 | 0.8579 | 0.8810 | 0.8923 | 0.8859 | 0.8840 | 0.8941 | 0.8884 | 0.9564 | 0.9677 | 0.9613 | 0.9231 | 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 | 0.6667 | 0.6923 | 0.6923 |
| 0.0016 | 8.0202 | 70000 | 0.2156 | 18.9871 | 0.8527 | 0.8636 | 0.8574 | 0.8808 | 0.8923 | 0.8858 | 0.8825 | 0.8938 | 0.8874 | 0.9544 | 0.9653 | 0.9591 | 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 | 0.6667 | 0.6667 | 0.6667 |
| 0.001 | 9.1659 | 80000 | 0.2248 | 18.9659 | 0.8610 | 0.8698 | 0.8647 | 0.8903 | 0.9000 | 0.8944 | 0.8933 | 0.9020 | 0.8969 | 0.9569 | 0.9673 | 0.9614 | 0.9231 | 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 | 0.4286 | 0.5 | 0.4615 |
| 0.001 | 10.3116 | 90000 | 0.2344 | 18.6692 | 0.8763 | 0.8880 | 0.8814 | 0.9011 | 0.9133 | 0.9063 | 0.9039 | 0.9152 | 0.9087 | 0.9558 | 0.9694 | 0.9618 | 0.9231 | 0.9286 | 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 | 0.5833 | 0.6923 | 0.6667 |
| 0.0008 | 11.4574 | 100000 | 0.2362 | 18.8811 | 0.8607 | 0.8709 | 0.8650 | 0.8878 | 0.8983 | 0.8922 | 0.8905 | 0.8994 | 0.8942 | 0.9534 | 0.9654 | 0.9586 | 0.9231 | 0.9231 | 0.9286 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.6429 | 0.6923 | 0.6923 |
| 0.0005 | 12.6031 | 110000 | 0.2331 | 18.3513 | 0.8603 | 0.8711 | 0.8648 | 0.8878 | 0.8992 | 0.8926 | 0.8897 | 0.9010 | 0.8945 | 0.9516 | 0.9631 | 0.9566 | 0.9231 | 0.9231 | 0.9286 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1667 | 0.2 | 0.1818 |
| 0.0002 | 13.7489 | 120000 | 0.2378 | 18.6692 | 0.8691 | 0.8790 | 0.8734 | 0.8960 | 0.9066 | 0.9006 | 0.8999 | 0.9102 | 0.9043 | 0.9583 | 0.9688 | 0.9628 | 0.9231 | 0.9231 | 0.9286 | 1.0 | 1.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.5455 | 0.5455 | 0.5455 |
| 0.0005 | 14.8946 | 130000 | 0.2428 | 18.3090 | 0.8702 | 0.8790 | 0.8739 | 0.8960 | 0.9053 | 0.8999 | 0.8997 | 0.9072 | 0.9027 | 0.9563 | 0.9671 | 0.9609 | 0.9231 | 0.9231 | 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 | 0.6923 | 0.6923 | 0.6923 |
| 0.0001 | 16.0403 | 140000 | 0.2476 | 18.0123 | 0.8701 | 0.8823 | 0.8754 | 0.8959 | 0.9085 | 0.9014 | 0.8993 | 0.9109 | 0.9043 | 0.9584 | 0.9715 | 0.9641 | 0.9231 | 0.9286 | 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 | 0.6923 | 0.6923 | 0.6923 |
| 0.0 | 17.1861 | 150000 | 0.2520 | 18.4361 | 0.8639 | 0.8777 | 0.8700 | 0.8910 | 0.9054 | 0.8974 | 0.8941 | 0.9079 | 0.9002 | 0.9590 | 0.9723 | 0.9648 | 0.9231 | 0.9286 | 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 | 0.6923 | 0.7778 | 0.7500 |
| 0.0 | 18.3318 | 160000 | 0.2503 | 18.1394 | 0.8682 | 0.8791 | 0.8729 | 0.8945 | 0.9061 | 0.8995 | 0.8980 | 0.9083 | 0.9024 | 0.9596 | 0.9715 | 0.9648 | 0.9286 | 0.9286 | 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 | 0.6923 | 0.75 | 0.7200 |
| 0.0 | 19.4775 | 170000 | 0.2474 | 17.8004 | 0.8607 | 0.8714 | 0.8653 | 0.8866 | 0.8977 | 0.8913 | 0.8900 | 0.9006 | 0.8945 | 0.9601 | 0.9701 | 0.9643 | 0.9286 | 0.9286 | 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 | 0.6364 | 0.6364 | 0.6364 |
| 0.0 | 20.6233 | 180000 | 0.2517 | 17.9275 | 0.8658 | 0.8766 | 0.8704 | 0.8928 | 0.9040 | 0.8976 | 0.8955 | 0.9057 | 0.8998 | 0.9604 | 0.9722 | 0.9655 | 0.9231 | 0.9286 | 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 | 0.6364 | 0.6364 | 0.6364 |
| 0.0 | 21.7690 | 190000 | 0.2553 | 17.6520 | 0.8667 | 0.8780 | 0.8716 | 0.8924 | 0.9041 | 0.8974 | 0.8950 | 0.9060 | 0.8997 | 0.9597 | 0.9718 | 0.9649 | 0.9231 | 0.9286 | 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 | 0.6364 | 0.6364 | 0.6364 |
| 0.0 | 22.9148 | 200000 | 0.2584 | 17.4613 | 0.8722 | 0.8838 | 0.8772 | 0.8981 | 0.9102 | 0.9033 | 0.9008 | 0.9121 | 0.9057 | 0.9601 | 0.9730 | 0.9657 | 0.9286 | 0.9333 | 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 | 0.6364 | 0.6364 | 0.6364 |
### Framework versions
- Transformers 4.41.2
- Pytorch 2.2.1
- Datasets 2.20.0
- Tokenizers 0.19.1
|