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_da1
results: []
w2v2-base-pretrained_lr5e-5_at0.8_da1
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: 5.2151
- Wer: 0.9633
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 |
|---|---|---|---|---|
| 16.3205 | 5.43 | 250 | 8.3965 | 1.0 |
| 4.5608 | 10.87 | 500 | 5.8173 | 1.0 |
| 3.4435 | 16.3 | 750 | 6.0815 | 1.0 |
| 3.2716 | 21.74 | 1000 | 5.0685 | 1.0 |
| 3.1606 | 27.17 | 1250 | 4.2348 | 1.0009 |
| 2.9691 | 32.61 | 1500 | 5.6780 | 1.0009 |
| 2.7899 | 38.04 | 1750 | 5.5722 | 1.0017 |
| 2.6205 | 43.48 | 2000 | 6.3790 | 1.0026 |
| 2.444 | 48.91 | 2250 | 5.4383 | 0.9919 |
| 2.3126 | 54.35 | 2500 | 5.0747 | 0.9940 |
| 2.0991 | 59.78 | 2750 | 5.3164 | 0.9855 |
| 1.8757 | 65.22 | 3000 | 5.7173 | 0.9833 |
| 1.7387 | 70.65 | 3250 | 5.4941 | 0.9940 |
| 1.6081 | 76.09 | 3500 | 5.3666 | 0.9675 |
| 1.53 | 81.52 | 3750 | 5.4622 | 0.9731 |
| 1.4346 | 86.96 | 4000 | 5.2151 | 0.9633 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.0.0
- Datasets 2.14.6
- Tokenizers 0.14.1