|
--- |
|
language: |
|
- sv-SE |
|
license: apache-2.0 |
|
tags: |
|
- automatic-speech-recognition |
|
- mozilla-foundation/common_voice_7_0 |
|
- generated_from_trainer |
|
datasets: |
|
- common_voice |
|
model-index: |
|
- name: '' |
|
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. --> |
|
|
|
# |
|
|
|
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - SV-SE dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.3179 |
|
- Wer: 0.2735 |
|
|
|
## 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: 7.5e-05 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- 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: 2000 |
|
- num_epochs: 50.0 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-----:|:-----:|:---------------:|:------:| |
|
| 3.3332 | 1.45 | 500 | 3.2920 | 1.0 | |
|
| 2.9269 | 2.91 | 1000 | 2.9415 | 0.9966 | |
|
| 2.0719 | 4.36 | 1500 | 1.1641 | 0.8508 | |
|
| 1.7404 | 5.81 | 2000 | 0.7281 | 0.6846 | |
|
| 1.5921 | 7.27 | 2500 | 0.5886 | 0.5147 | |
|
| 1.4941 | 8.72 | 3000 | 0.5183 | 0.5063 | |
|
| 1.4486 | 10.17 | 3500 | 0.4749 | 0.4676 | |
|
| 1.3899 | 11.63 | 4000 | 0.4565 | 0.4432 | |
|
| 1.3881 | 13.08 | 4500 | 0.4316 | 0.4228 | |
|
| 1.3572 | 14.53 | 5000 | 0.4195 | 0.3834 | |
|
| 1.3261 | 15.99 | 5500 | 0.3974 | 0.3607 | |
|
| 1.2809 | 17.44 | 6000 | 0.3845 | 0.3467 | |
|
| 1.2713 | 18.89 | 6500 | 0.3832 | 0.3450 | |
|
| 1.257 | 20.35 | 7000 | 0.3779 | 0.3373 | |
|
| 1.2298 | 21.8 | 7500 | 0.3744 | 0.3391 | |
|
| 1.2173 | 23.26 | 8000 | 0.3745 | 0.3262 | |
|
| 1.1966 | 24.71 | 8500 | 0.3680 | 0.3241 | |
|
| 1.1925 | 26.16 | 9000 | 0.3605 | 0.3171 | |
|
| 1.1692 | 27.61 | 9500 | 0.3512 | 0.3147 | |
|
| 1.1704 | 29.07 | 10000 | 0.3532 | 0.3098 | |
|
| 1.1595 | 30.52 | 10500 | 0.3425 | 0.3039 | |
|
| 1.1433 | 31.97 | 11000 | 0.3568 | 0.3026 | |
|
| 1.1295 | 33.43 | 11500 | 0.3461 | 0.2992 | |
|
| 1.1131 | 34.88 | 12000 | 0.3349 | 0.2942 | |
|
| 1.1015 | 36.34 | 12500 | 0.3378 | 0.2961 | |
|
| 1.0835 | 37.79 | 13000 | 0.3282 | 0.2865 | |
|
| 1.083 | 39.24 | 13500 | 0.3182 | 0.2826 | |
|
| 1.0819 | 40.7 | 14000 | 0.3264 | 0.2850 | |
|
| 1.072 | 42.15 | 14500 | 0.3279 | 0.2817 | |
|
| 1.0456 | 43.6 | 15000 | 0.3234 | 0.2793 | |
|
| 1.0581 | 45.06 | 15500 | 0.3220 | 0.2779 | |
|
| 1.0406 | 46.51 | 16000 | 0.3208 | 0.2762 | |
|
| 1.0422 | 47.96 | 16500 | 0.3184 | 0.2752 | |
|
| 1.0099 | 49.42 | 17000 | 0.3181 | 0.2735 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.16.0.dev0 |
|
- Pytorch 1.10.1+cu102 |
|
- Datasets 1.17.1.dev0 |
|
- Tokenizers 0.11.0 |
|
|
