|
--- |
|
license: apache-2.0 |
|
base_model: facebook/wav2vec2-large-xlsr-53 |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: wav2vec2-timit-xls-r-53-wandb-colab |
|
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. --> |
|
|
|
# wav2vec2-timit-xls-r-53-wandb-colab |
|
|
|
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: 0.2451 |
|
- Wer: 0.2503 |
|
- Cer: 0.0799 |
|
|
|
## 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: 0.0001 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 1000 |
|
- num_epochs: 30 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------:|:------:| |
|
| No log | 0.69 | 400 | 3.1514 | 1.0 | 0.9790 | |
|
| 5.094 | 1.38 | 800 | 2.8674 | 1.0 | 0.9790 | |
|
| 2.595 | 2.08 | 1200 | 0.5208 | 0.5344 | 0.1541 | |
|
| 0.8139 | 2.77 | 1600 | 0.3568 | 0.4234 | 0.1275 | |
|
| 0.4793 | 3.46 | 2000 | 0.2954 | 0.3645 | 0.1106 | |
|
| 0.4793 | 4.15 | 2400 | 0.2649 | 0.3475 | 0.1037 | |
|
| 0.3771 | 4.84 | 2800 | 0.2452 | 0.3186 | 0.0976 | |
|
| 0.2912 | 5.54 | 3200 | 0.2385 | 0.3079 | 0.0960 | |
|
| 0.2632 | 6.23 | 3600 | 0.2292 | 0.2954 | 0.0911 | |
|
| 0.2176 | 6.92 | 4000 | 0.2248 | 0.2910 | 0.0908 | |
|
| 0.2176 | 7.61 | 4400 | 0.2279 | 0.2816 | 0.0888 | |
|
| 0.1958 | 8.3 | 4800 | 0.2227 | 0.2819 | 0.0878 | |
|
| 0.1846 | 9.0 | 5200 | 0.2277 | 0.2779 | 0.0876 | |
|
| 0.1573 | 9.69 | 5600 | 0.2280 | 0.2830 | 0.0877 | |
|
| 0.1471 | 10.38 | 6000 | 0.2345 | 0.2770 | 0.0880 | |
|
| 0.1471 | 11.07 | 6400 | 0.2389 | 0.2714 | 0.0852 | |
|
| 0.133 | 11.76 | 6800 | 0.2253 | 0.2730 | 0.0869 | |
|
| 0.1317 | 12.46 | 7200 | 0.2179 | 0.2662 | 0.0846 | |
|
| 0.1268 | 13.15 | 7600 | 0.2315 | 0.2678 | 0.0851 | |
|
| 0.1147 | 13.84 | 8000 | 0.2501 | 0.2679 | 0.0849 | |
|
| 0.1147 | 14.53 | 8400 | 0.2463 | 0.2663 | 0.0839 | |
|
| 0.1151 | 15.22 | 8800 | 0.2429 | 0.2662 | 0.0848 | |
|
| 0.0968 | 15.92 | 9200 | 0.2502 | 0.2639 | 0.0839 | |
|
| 0.0985 | 16.61 | 9600 | 0.2589 | 0.2616 | 0.0838 | |
|
| 0.0938 | 17.3 | 10000 | 0.2414 | 0.2595 | 0.0835 | |
|
| 0.0938 | 17.99 | 10400 | 0.2420 | 0.2617 | 0.0839 | |
|
| 0.0878 | 18.69 | 10800 | 0.2257 | 0.2597 | 0.0829 | |
|
| 0.0872 | 19.38 | 11200 | 0.2654 | 0.2586 | 0.0825 | |
|
| 0.0774 | 20.07 | 11600 | 0.2558 | 0.2579 | 0.0829 | |
|
| 0.0743 | 20.76 | 12000 | 0.2375 | 0.2564 | 0.0824 | |
|
| 0.0743 | 21.45 | 12400 | 0.2522 | 0.2568 | 0.0813 | |
|
| 0.0832 | 22.15 | 12800 | 0.2363 | 0.2569 | 0.0817 | |
|
| 0.0698 | 22.84 | 13200 | 0.2510 | 0.2574 | 0.0816 | |
|
| 0.0677 | 23.53 | 13600 | 0.2535 | 0.2570 | 0.0818 | |
|
| 0.0648 | 24.22 | 14000 | 0.2595 | 0.2571 | 0.0819 | |
|
| 0.0648 | 24.91 | 14400 | 0.2441 | 0.2542 | 0.0815 | |
|
| 0.0685 | 25.61 | 14800 | 0.2503 | 0.2534 | 0.0803 | |
|
| 0.066 | 26.3 | 15200 | 0.2489 | 0.2533 | 0.0804 | |
|
| 0.0583 | 26.99 | 15600 | 0.2471 | 0.2512 | 0.0802 | |
|
| 0.0624 | 27.68 | 16000 | 0.2487 | 0.2516 | 0.0804 | |
|
| 0.0624 | 28.37 | 16400 | 0.2470 | 0.2518 | 0.0804 | |
|
| 0.0663 | 29.07 | 16800 | 0.2478 | 0.2510 | 0.0800 | |
|
| 0.0569 | 29.76 | 17200 | 0.2451 | 0.2503 | 0.0799 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.32.0.dev0 |
|
- Pytorch 2.0.1+cu118 |
|
- Datasets 1.18.3 |
|
- Tokenizers 0.13.3 |
|
|
