<!-- 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-xls-r-300m-arabic_speech_commands_half
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: 0.2678
- Accuracy: 0.9975
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: 32
- eval_batch_size: 32
- seed: 42
- 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: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 3.6816 | 0.99 | 28 | 3.4693 | 0.1575 |
| 3.0227 | 1.99 | 56 | 2.5330 | 0.2775 |
| 2.3345 | 2.99 | 84 | 1.8723 | 0.5925 |
| 1.6785 | 3.99 | 112 | 1.2944 | 0.8092 |
| 1.2606 | 4.99 | 140 | 0.8933 | 0.9425 |
| 1.0024 | 5.99 | 168 | 0.6041 | 0.9817 |
| 0.6478 | 6.99 | 196 | 0.3814 | 0.9933 |
| 0.4768 | 7.99 | 224 | 0.2678 | 0.9975 |
| 0.4143 | 8.99 | 252 | 0.2198 | 0.9967 |
| 0.3278 | 9.99 | 280 | 0.1993 | 0.9967 |
Framework versions
- Transformers 4.21.1
- Pytorch 1.12.1
- Datasets 2.4.0
- Tokenizers 0.12.1
