<!-- 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. -->
nypnop-wav2vec2-large-xlsr-indo-demo-colab
This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the common_voice_11_0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4002
- Wer: 0.4282
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: 16
- eval_batch_size: 8
- 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: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer |
---|---|---|---|---|
4.9825 | 2.92 | 400 | 2.7861 | 1.0000 |
1.2787 | 5.84 | 800 | 0.4612 | 0.6220 |
0.3814 | 8.76 | 1200 | 0.4132 | 0.5438 |
0.2646 | 11.68 | 1600 | 0.4015 | 0.4914 |
0.2098 | 14.6 | 2000 | 0.3914 | 0.4806 |
0.1681 | 17.52 | 2400 | 0.3934 | 0.4627 |
0.1522 | 20.44 | 2800 | 0.3977 | 0.4498 |
0.1294 | 23.36 | 3200 | 0.4121 | 0.4398 |
0.1197 | 26.28 | 3600 | 0.4119 | 0.4291 |
0.1084 | 29.2 | 4000 | 0.4002 | 0.4282 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.0.0+cu118
- Datasets 2.6.1
- Tokenizers 0.13.3