metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: w2v2-base-pretrained_lr5e-5_at1_da1
results: []
w2v2-base-pretrained_lr5e-5_at1_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: 2.2991
- Wer: 0.3016
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 |
|---|---|---|---|---|
| 17.3296 | 6.1 | 250 | 8.4283 | 1.0 |
| 4.6394 | 12.2 | 500 | 7.5283 | 1.0 |
| 3.4235 | 18.29 | 750 | 5.8716 | 1.0 |
| 3.3011 | 24.39 | 1000 | 4.9842 | 1.0 |
| 3.048 | 30.49 | 1250 | 3.1389 | 1.0077 |
| 2.1291 | 36.59 | 1500 | 2.0209 | 0.9257 |
| 0.5553 | 42.68 | 1750 | 1.5785 | 0.4742 |
| 0.2117 | 48.78 | 2000 | 1.7978 | 0.3631 |
| 0.1228 | 54.88 | 2250 | 1.8723 | 0.3323 |
| 0.0771 | 60.98 | 2500 | 2.0055 | 0.3234 |
| 0.0587 | 67.07 | 2750 | 2.0493 | 0.3157 |
| 0.0398 | 73.17 | 3000 | 2.0708 | 0.3084 |
| 0.0336 | 79.27 | 3250 | 2.2532 | 0.3148 |
| 0.0271 | 85.37 | 3500 | 2.2694 | 0.3016 |
| 0.0234 | 91.46 | 3750 | 2.3538 | 0.3050 |
| 0.0213 | 97.56 | 4000 | 2.2991 | 0.3016 |
Framework versions
- Transformers 4.35.0
- Pytorch 2.0.0
- Datasets 2.14.6
- Tokenizers 0.14.1