metadata
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1
results: []
ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1
This model is a fine-tuned version of gary109/ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.7044
- Wer: 0.4757
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: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 10.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.0489 | 1.0 | 72 | 0.9702 | 0.6696 |
| 0.9699 | 2.0 | 144 | 0.9321 | 0.6596 |
| 0.9074 | 3.0 | 216 | 0.8542 | 0.6139 |
| 0.8979 | 4.0 | 288 | 0.8187 | 0.5777 |
| 0.8363 | 5.0 | 360 | 0.7735 | 0.5551 |
| 0.802 | 6.0 | 432 | 0.7766 | 0.5402 |
| 0.7702 | 7.0 | 504 | 0.7471 | 0.4971 |
| 0.7483 | 8.0 | 576 | 0.7226 | 0.4981 |
| 0.7196 | 9.0 | 648 | 0.7095 | 0.4842 |
| 0.731 | 10.0 | 720 | 0.7044 | 0.4757 |
Framework versions
- Transformers 4.21.0.dev0
- Pytorch 1.9.1+cu102
- Datasets 2.3.3.dev0
- Tokenizers 0.12.1