<!-- 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-base-50k
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice dataset. It achieves the following results on the evaluation set:
- Loss: 3.5640
- Wer: 1.0
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: 7.5e-05
- train_batch_size: 16
- eval_batch_size: 16
- 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: 1000
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
10.7005 | 0.48 | 300 | 5.3021 | 1.0 |
3.9938 | 0.96 | 600 | 3.4997 | 1.0 |
3.591 | 1.44 | 900 | 3.5641 | 1.0 |
3.6168 | 1.92 | 1200 | 3.5641 | 1.0 |
3.6252 | 2.4 | 1500 | 3.5641 | 1.0 |
3.6137 | 2.88 | 1800 | 3.5641 | 1.0 |
3.6124 | 3.36 | 2100 | 3.5641 | 1.0 |
3.6171 | 3.84 | 2400 | 3.5641 | 1.0 |
3.6436 | 4.32 | 2700 | 3.5641 | 1.0 |
3.6189 | 4.8 | 3000 | 3.5640 | 1.0 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.11.0
- Datasets 2.1.0
- Tokenizers 0.12.1