|
--- |
|
license: apache-2.0 |
|
base_model: facebook/wav2vec2-xls-r-300m |
|
tags: |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: wav2vec2_XLSR_darija_maroc |
|
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_XLSR_darija_maroc |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.2860 |
|
- Wer: 0.3290 |
|
|
|
## 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: 0.0003 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 32 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 500 |
|
- num_epochs: 30 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------:| |
|
| 4.9354 | 0.83 | 400 | 2.0492 | 1.0371 | |
|
| 0.8236 | 1.66 | 800 | 0.4434 | 0.5832 | |
|
| 0.4821 | 2.49 | 1200 | 0.3597 | 0.5114 | |
|
| 0.3823 | 3.32 | 1600 | 0.3265 | 0.4758 | |
|
| 0.3231 | 4.15 | 2000 | 0.3149 | 0.4526 | |
|
| 0.2854 | 4.97 | 2400 | 0.2797 | 0.4237 | |
|
| 0.2529 | 5.8 | 2800 | 0.3027 | 0.4415 | |
|
| 0.2493 | 6.63 | 3200 | 0.2926 | 0.4264 | |
|
| 0.2138 | 7.46 | 3600 | 0.2857 | 0.4169 | |
|
| 0.2067 | 8.29 | 4000 | 0.2743 | 0.4099 | |
|
| 0.1898 | 9.12 | 4400 | 0.2798 | 0.3993 | |
|
| 0.1755 | 9.95 | 4800 | 0.2800 | 0.3913 | |
|
| 0.1603 | 10.78 | 5200 | 0.2709 | 0.3860 | |
|
| 0.1608 | 11.61 | 5600 | 0.2716 | 0.3872 | |
|
| 0.1462 | 12.44 | 6000 | 0.2697 | 0.3825 | |
|
| 0.137 | 13.26 | 6400 | 0.2855 | 0.3819 | |
|
| 0.1326 | 14.09 | 6800 | 0.2860 | 0.3733 | |
|
| 0.123 | 14.92 | 7200 | 0.2677 | 0.3813 | |
|
| 0.1168 | 15.75 | 7600 | 0.2780 | 0.3740 | |
|
| 0.1113 | 16.58 | 8000 | 0.2926 | 0.3719 | |
|
| 0.1057 | 17.41 | 8400 | 0.2927 | 0.3704 | |
|
| 0.0996 | 18.24 | 8800 | 0.2825 | 0.3602 | |
|
| 0.0967 | 19.07 | 9200 | 0.2983 | 0.3641 | |
|
| 0.0925 | 19.9 | 9600 | 0.2843 | 0.3576 | |
|
| 0.0894 | 20.73 | 10000 | 0.2726 | 0.3668 | |
|
| 0.0836 | 21.55 | 10400 | 0.2829 | 0.3560 | |
|
| 0.0789 | 22.38 | 10800 | 0.2806 | 0.3508 | |
|
| 0.0778 | 23.21 | 11200 | 0.2849 | 0.3540 | |
|
| 0.0742 | 24.04 | 11600 | 0.2770 | 0.3436 | |
|
| 0.0679 | 24.87 | 12000 | 0.2850 | 0.3425 | |
|
| 0.063 | 25.7 | 12400 | 0.2846 | 0.3366 | |
|
| 0.0593 | 26.53 | 12800 | 0.2811 | 0.3351 | |
|
| 0.0586 | 27.36 | 13200 | 0.2863 | 0.3322 | |
|
| 0.0555 | 28.19 | 13600 | 0.2819 | 0.3311 | |
|
| 0.053 | 29.02 | 14000 | 0.2874 | 0.3301 | |
|
| 0.0498 | 29.84 | 14400 | 0.2860 | 0.3290 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.34.1 |
|
- Pytorch 2.1.0+cu118 |
|
- Datasets 2.14.6 |
|
- Tokenizers 0.14.1 |
|
|
