wav2vec-large-xls-r-300-ha-colab_4
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_10_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.8545
- Wer: 0.5860
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: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- 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: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.8309 | 5.19 | 400 | 0.9834 | 0.8145 |
| 0.3782 | 10.39 | 800 | 0.7377 | 0.6526 |
| 0.1718 | 15.58 | 1200 | 0.7885 | 0.6293 |
| 0.1017 | 20.77 | 1600 | 0.9308 | 0.6115 |
| 0.0664 | 25.97 | 2000 | 0.8545 | 0.5860 |
Framework versions
- Transformers 4.11.3
- Pytorch 1.10.0+cu113
- Datasets 2.3.2
- Tokenizers 0.10.3
- Downloads last month
- 9
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.