metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: base
results: []
base
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: 1.7496
- Wer: 45.8524
- Cer: 23.3228
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: 0.0001
- train_batch_size: 32
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2000
- training_steps: 5000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 3.3189 | 18.52 | 1000 | 1.3203 | 61.7505 | 28.6087 |
| 0.4697 | 37.04 | 2000 | 1.3792 | 48.5805 | 24.3342 |
| 0.1681 | 55.56 | 3000 | 1.4859 | 47.0433 | 23.6677 |
| 0.0913 | 74.07 | 4000 | 1.6549 | 46.4617 | 23.4416 |
| 0.0598 | 92.59 | 5000 | 1.7496 | 45.8524 | 23.3228 |
Framework versions
- Transformers 4.39.3
- Pytorch 1.12.1+cu116
- Datasets 2.18.0
- Tokenizers 0.15.2