metadata
license: apache-2.0
base_model: facebook/wav2vec2-xls-r-300m
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2-xls-r-300m-MCV15
results: []
datasets:
- mozilla-foundation/common_voice_15_0
language:
- br
wav2vec2-xls-r-300m-MCV15
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.1155
- Wer: 0.6060
- Cer: 0.2242
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: 8e-05
- train_batch_size: 24
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 48
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 60
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 9.8843 | 4.5 | 250 | 3.3243 | 1.0 | 1.0000 |
| 3.066 | 9.01 | 500 | 2.8141 | 1.0 | 1.0000 |
| 1.8006 | 13.51 | 750 | 1.1647 | 0.7936 | 0.3028 |
| 0.8925 | 18.02 | 1000 | 1.0696 | 0.7174 | 0.2679 |
| 0.6306 | 22.52 | 1250 | 1.0330 | 0.6792 | 0.2509 |
| 0.512 | 27.03 | 1500 | 0.9988 | 0.6560 | 0.2405 |
| 0.4275 | 31.53 | 1750 | 1.0428 | 0.6345 | 0.2332 |
| 0.3837 | 36.04 | 2000 | 1.0555 | 0.6267 | 0.2315 |
| 0.3481 | 40.54 | 2250 | 1.1165 | 0.6231 | 0.2312 |
| 0.3081 | 45.05 | 2500 | 1.0772 | 0.6142 | 0.2255 |
| 0.2906 | 49.55 | 2750 | 1.1146 | 0.6085 | 0.2267 |
| 0.2923 | 54.05 | 3000 | 1.1030 | 0.6058 | 0.2228 |
| 0.259 | 58.56 | 3250 | 1.1155 | 0.6060 | 0.2242 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.0
- Tokenizers 0.15.0