metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: wav2vec2_batangueno
results: []
wav2vec2_batangueno
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: 0.5121
- Wer: 0.2514
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: 6.642954074604246e-05
- train_batch_size: 2
- 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: 1000
- num_epochs: 30
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.8266 | 2.67 | 500 | 2.8513 | 1.0 |
| 1.587 | 5.35 | 1000 | 0.8559 | 0.5943 |
| 0.5172 | 8.02 | 1500 | 0.5303 | 0.3903 |
| 0.2865 | 10.7 | 2000 | 0.5031 | 0.3716 |
| 0.2002 | 13.37 | 2500 | 0.4735 | 0.2800 |
| 0.1566 | 16.04 | 3000 | 0.4878 | 0.2955 |
| 0.1163 | 18.72 | 3500 | 0.5329 | 0.2889 |
| 0.0948 | 21.39 | 4000 | 0.4727 | 0.2679 |
| 0.0797 | 24.06 | 4500 | 0.4704 | 0.2679 |
| 0.0669 | 26.74 | 5000 | 0.5199 | 0.2580 |
| 0.0579 | 29.41 | 5500 | 0.5121 | 0.2514 |
Framework versions
- Transformers 4.34.0
- Pytorch 2.0.1+cu118
- Datasets 1.18.3
- Tokenizers 0.14.1