metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- google/fleurs
metrics:
- wer
model-index:
- name: xls-r-fleurs_zu-run1
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: audiofolder
type: audiofolder
config: default
split: validation
args: default
metrics:
- name: Wer
type: wer
value: 0.600381
xls-r-fleurs_zu-run1
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the FLEURS (zu) dataset. It achieves the following results:
- Wer (Validation): 61.41%
- Wer (Test): 60.23%
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 12
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer (Train) |
|---|---|---|---|---|
| 50 | 19.607200 | 9.162900 | 1.000000 | |
| 100 | 7.038300 | 4.738822 | 1.000000 | |
| 150 | 4.190100 | 3.359574 | 1.000000 | |
| 200 | 3.161900 | 3.032595 | 1.000000 | |
| 250 | 3.004700 | 2.994741 | 1.000000 | |
| 300 | 2.988300 | 2.955285 | 1.000000 | |
| 350 | 2.675800 | 1.816109 | 1.000000 | |
| 400 | 1.064400 | 0.866473 | 0.814220 | |
| 450 | 0.601600 | 0.696754 | 0.712340 | |
| 500 | 0.506900 | 0.662974 | 0.716426 | |
| 550 | 0.432200 | 0.598446 | 0.667121 | |
| 600 | 0.358700 | 0.618853 | 0.681013 | |
| 650 | 0.333300 | 0.564290 | 0.627349 | |
| 700 | 0.283100 | 0.573746 | 0.646418 | |
| 750 | 0.250800 | 0.577737 | 0.639608 | |
| 800 | 0.232200 | 0.557288 | 0.604467 | |
| 850 | 0.191200 | 0.538959 | 0.590030 | |
| 900 | 0.195600 | 0.549700 | 0.600654 | |
| 950 | 0.193000 | 0.579098 | 0.611278 | |
| 1000 | 0.169900 | 0.578752 | 0.600381 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3