<!-- 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-ft-keyword-spotting
This model is a fine-tuned version of facebook/wav2vec2-base on the superb dataset. It achieves the following results on the evaluation set:
- Loss: 0.0795
- Accuracy: 0.9829
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: 0
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.5546 | 1.0 | 399 | 0.4250 | 0.9618 |
| 0.2128 | 2.0 | 798 | 0.1331 | 0.9781 |
| 0.1763 | 3.0 | 1197 | 0.0935 | 0.9807 |
| 0.1485 | 4.0 | 1596 | 0.0852 | 0.9828 |
| 0.1335 | 5.0 | 1995 | 0.0795 | 0.9829 |
Framework versions
- Transformers 4.25.1
- Pytorch 1.10.0+cu111
- Datasets 2.7.1
- Tokenizers 0.13.2
