metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- common_voice_8_0
metrics:
- wer
model-index:
- name: wav2vec2-large-xls-r-1b-frisian-cv-8-10m
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: common_voice_8_0
type: common_voice_8_0
config: fy-NL
split: validation
args: fy-NL
metrics:
- name: Wer
type: wer
value: 0.8409155834880732
wav2vec2-large-xls-r-1b-frisian-cv-8-10m
This model is a fine-tuned version of facebook/wav2vec2-xls-r-1b on the common_voice_8_0 dataset. It achieves the following results on the evaluation set:
- Loss: 1.2060
- Wer: 0.8409
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: 7e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.98) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 90
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 13.1755 | 6.25 | 50 | 4.7246 | 0.9927 |
| 8.0106 | 12.5 | 100 | 3.8216 | 0.9996 |
| 3.4536 | 18.75 | 150 | 3.0254 | 1.0 |
| 3.1828 | 25.0 | 200 | 2.9839 | 1.0 |
| 3.1087 | 31.25 | 250 | 2.9887 | 1.0 |
| 3.0176 | 37.5 | 300 | 2.8903 | 1.0 |
| 2.912 | 43.75 | 350 | 2.6903 | 1.0 |
| 2.7648 | 50.0 | 400 | 2.1968 | 1.0016 |
| 2.1665 | 56.25 | 450 | 1.6039 | 0.9838 |
| 1.8149 | 62.5 | 500 | 1.3367 | 0.9484 |
| 1.6385 | 68.75 | 550 | 1.2353 | 0.9340 |
| 1.314 | 75.0 | 600 | 1.2081 | 0.8834 |
| 1.0963 | 81.25 | 650 | 1.2016 | 0.8416 |
| 1.2322 | 87.5 | 700 | 1.2060 | 0.8409 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu117
- Datasets 2.11.0
- Tokenizers 0.13.3