|
--- |
|
license: apache-2.0 |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: kids_phoneme_sm_model |
|
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. --> |
|
|
|
# kids_phoneme_sm_model |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.6936 |
|
- Cer: 0.2531 |
|
|
|
## 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: 4e-05 |
|
- train_batch_size: 2 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 30 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Cer | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------:| |
|
| 3.2437 | 0.74 | 500 | 4.1235 | 1.0 | |
|
| 2.8562 | 1.48 | 1000 | 3.5824 | 1.0 | |
|
| 2.7606 | 2.22 | 1500 | 3.2239 | 1.0 | |
|
| 2.0885 | 2.96 | 2000 | 1.1613 | 0.8147 | |
|
| 1.0295 | 3.7 | 2500 | 0.7703 | 0.5125 | |
|
| 0.796 | 4.44 | 3000 | 0.6539 | 0.4420 | |
|
| 0.6484 | 5.19 | 3500 | 0.6259 | 0.3937 | |
|
| 0.6099 | 5.93 | 4000 | 0.5749 | 0.3887 | |
|
| 0.5772 | 6.67 | 4500 | 0.6031 | 0.3637 | |
|
| 0.5158 | 7.41 | 5000 | 0.5978 | 0.3518 | |
|
| 0.4923 | 8.15 | 5500 | 0.5621 | 0.3364 | |
|
| 0.4679 | 8.89 | 6000 | 0.5371 | 0.3396 | |
|
| 0.4385 | 9.63 | 6500 | 0.5804 | 0.3213 | |
|
| 0.4818 | 10.37 | 7000 | 0.5469 | 0.3223 | |
|
| 0.3797 | 11.11 | 7500 | 0.5789 | 0.3118 | |
|
| 0.3669 | 11.85 | 8000 | 0.5733 | 0.2986 | |
|
| 0.3777 | 12.59 | 8500 | 0.6053 | 0.3004 | |
|
| 0.3613 | 13.33 | 9000 | 0.6061 | 0.2895 | |
|
| 0.3454 | 14.07 | 9500 | 0.6072 | 0.2740 | |
|
| 0.3532 | 14.81 | 10000 | 0.6119 | 0.2872 | |
|
| 0.3087 | 15.56 | 10500 | 0.6020 | 0.2849 | |
|
| 0.3277 | 16.3 | 11000 | 0.6397 | 0.2745 | |
|
| 0.2978 | 17.04 | 11500 | 0.6216 | 0.2745 | |
|
| 0.2939 | 17.78 | 12000 | 0.6377 | 0.2690 | |
|
| 0.2675 | 18.52 | 12500 | 0.6752 | 0.2681 | |
|
| 0.2873 | 19.26 | 13000 | 0.6677 | 0.2767 | |
|
| 0.2779 | 20.0 | 13500 | 0.6748 | 0.2717 | |
|
| 0.28 | 20.74 | 14000 | 0.6771 | 0.2645 | |
|
| 0.2688 | 21.48 | 14500 | 0.6618 | 0.2604 | |
|
| 0.2234 | 22.22 | 15000 | 0.6791 | 0.2613 | |
|
| 0.2464 | 22.96 | 15500 | 0.6665 | 0.2626 | |
|
| 0.2254 | 23.7 | 16000 | 0.7028 | 0.2572 | |
|
| 0.2132 | 24.44 | 16500 | 0.6985 | 0.2567 | |
|
| 0.2424 | 25.19 | 17000 | 0.6731 | 0.2590 | |
|
| 0.2447 | 25.93 | 17500 | 0.6780 | 0.2544 | |
|
| 0.2209 | 26.67 | 18000 | 0.6729 | 0.2567 | |
|
| 0.2102 | 27.41 | 18500 | 0.6844 | 0.2563 | |
|
| 0.2185 | 28.15 | 19000 | 0.6922 | 0.2585 | |
|
| 0.2294 | 28.89 | 19500 | 0.6940 | 0.2563 | |
|
| 0.2208 | 29.63 | 20000 | 0.6936 | 0.2531 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.30.1 |
|
- Pytorch 2.0.0 |
|
- Datasets 2.12.0 |
|
- Tokenizers 0.13.3 |
|
|
