metadata
library_name: transformers
license: mit
base_model: facebook/w2v-bert-2.0
tags:
- automatic-speech-recognition
- CLEAR-Global/luo_19h
- generated_from_trainer
metrics:
- wer
model-index:
- name: w2v-bert-2.0-luo_cv_fleurs_19h
results: []
w2v-bert-2.0-luo_cv_fleurs_19h
This model is a fine-tuned version of facebook/w2v-bert-2.0 on the CLEAR-GLOBAL/LUO_19H - NA dataset. It achieves the following results on the evaluation set:
- Loss: 0.2682
- Wer: 0.2998
- Cer: 0.0930
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: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 100000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.698 | 6.4935 | 1000 | 0.7171 | 0.5988 | 0.1884 |
| 0.2666 | 12.9870 | 2000 | 0.3521 | 0.3862 | 0.1107 |
| 0.1497 | 19.4805 | 3000 | 0.2914 | 0.3351 | 0.0979 |
| 0.0802 | 25.9740 | 4000 | 0.2682 | 0.2976 | 0.0931 |
| 0.053 | 32.4675 | 5000 | 0.3036 | 0.3060 | 0.0913 |
| 0.0309 | 38.9610 | 6000 | 0.3689 | 0.2906 | 0.0939 |
| 0.0245 | 45.4545 | 7000 | 0.4164 | 0.3792 | 0.1007 |
| 0.0122 | 51.9481 | 8000 | 0.3996 | 0.3166 | 0.0964 |
| 0.0088 | 58.4416 | 9000 | 0.4323 | 0.3056 | 0.0952 |
Framework versions
- Transformers 4.48.1
- Pytorch 2.6.0+cu124
- Datasets 3.5.0
- Tokenizers 0.21.1