<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
wav2vec2-large-xls-r-300m-dsb-with-hsb-pretraining
This model is a fine-tuned version of TiMauzi/wav2vec2-large-xls-r-300m-hsb on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.3006
- Wer: 0.4336
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: 12
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 3
- total_train_batch_size: 36
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.6035 | 2.87 | 200 | 1.1605 | 0.5424 |
| 0.826 | 5.74 | 400 | 1.2218 | 0.5430 |
| 0.5227 | 8.61 | 600 | 1.3488 | 0.4680 |
| 0.4012 | 11.48 | 800 | 1.1785 | 0.4430 |
| 0.2807 | 14.35 | 1000 | 1.4286 | 0.4384 |
| 0.2167 | 17.22 | 1200 | 1.3006 | 0.4336 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
