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_da0.025
results: []
w2v2-base-pretrained_lr5e-5_at0.8_da0.025
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: 9.1596
- Wer: 1.0
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 |
|---|---|---|---|---|
| 33.8202 | 125.0 | 250 | 47.1503 | 1.0 |
| 11.5918 | 250.0 | 500 | 9.7965 | 1.0 |
| 10.0921 | 375.0 | 750 | 9.2904 | 1.0 |
| 6.6014 | 500.0 | 1000 | 9.2396 | 1.0 |
| 7.8557 | 625.0 | 1250 | 9.2582 | 1.0 |
| 7.6576 | 750.0 | 1500 | 10.7610 | 1.0 |
| 7.0009 | 875.0 | 1750 | 9.2900 | 1.0 |
| 6.6034 | 1000.0 | 2000 | 9.2763 | 1.0 |
| 6.5928 | 1125.0 | 2250 | 9.2703 | 1.0 |
| 6.5581 | 1250.0 | 2500 | 9.1374 | 1.0 |
| 6.4663 | 1375.0 | 2750 | 9.1508 | 1.0 |
| 6.4496 | 1500.0 | 3000 | 9.1802 | 1.0 |
| 6.4061 | 1625.0 | 3250 | 9.1858 | 1.0 |
| 6.4291 | 1750.0 | 3500 | 9.1535 | 1.0 |
| 6.4131 | 1875.0 | 3750 | 9.1450 | 1.0 |
| 6.4086 | 2000.0 | 4000 | 9.1596 | 1.0 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.0.0
- Datasets 2.14.6
- Tokenizers 0.14.1