<!-- 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-Arabic-phonemeIPA
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: 0.0398
- Per: 0.0833
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: 0.0005
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- 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: 250
- num_epochs: 9
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Per |
---|---|---|---|---|
0.1294 | 1.0 | 102 | 0.0673 | 0.1060 |
0.1964 | 2.0 | 204 | 0.0753 | 0.1167 |
0.2395 | 3.0 | 306 | 0.0851 | 0.1134 |
0.2317 | 4.0 | 408 | 0.0849 | 0.1152 |
0.2082 | 5.0 | 510 | 0.0853 | 0.1085 |
0.1856 | 6.0 | 612 | 0.0626 | 0.0946 |
0.1616 | 7.0 | 714 | 0.0635 | 0.0892 |
0.1426 | 8.0 | 816 | 0.0554 | 0.0863 |
0.1284 | 9.0 | 918 | 0.0555 | 0.0846 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1+cu116
- Datasets 2.10.1
- Tokenizers 0.13.2