metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: w2v2-base-pretrained_lr5e-5_at0.8_da0.05
results: []
w2v2-base-pretrained_lr5e-5_at0.8_da0.05
This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:
- Loss: 4.9499
- Wer: 1.1098
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: 32
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 32.094 | 83.33 | 250 | 3.6861 | 1.0 |
| 3.223 | 166.67 | 500 | 3.1064 | 1.0 |
| 3.0141 | 250.0 | 750 | 3.0722 | 1.0 |
| 2.6901 | 333.33 | 1000 | 2.8675 | 1.0 |
| 1.7283 | 416.67 | 1250 | 3.2044 | 1.0436 |
| 0.9791 | 500.0 | 1500 | 3.4905 | 1.0666 |
| 0.6506 | 583.33 | 1750 | 3.7132 | 1.0560 |
| 0.3895 | 666.67 | 2000 | 4.1587 | 1.0824 |
| 0.2399 | 750.0 | 2250 | 4.4319 | 1.1012 |
| 0.1794 | 833.33 | 2500 | 4.4773 | 1.0995 |
| 0.139 | 916.67 | 2750 | 4.6309 | 1.0871 |
| 0.1026 | 1000.0 | 3000 | 4.6925 | 1.0628 |
| 0.0807 | 1083.33 | 3250 | 4.8052 | 1.0824 |
| 0.0611 | 1166.67 | 3500 | 4.8086 | 1.0953 |
| 0.0539 | 1250.0 | 3750 | 4.8570 | 1.0995 |
| 0.0524 | 1333.33 | 4000 | 4.9499 | 1.1098 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.0.0
- Datasets 2.14.6
- Tokenizers 0.14.1