metadata
license: apache-2.0
tags:
- generated_from_trainer
datasets:
- null
model_index:
- name: wav2vec2-base-timit-demo-colab
results:
- {}
wav2vec2-base-timit-demo-colab
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: 0.4634
- Wer: 0.3469
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: 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
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 3.4469 | 4.0 | 500 | 1.2851 | 0.9547 |
| 0.616 | 8.0 | 1000 | 0.4867 | 0.4572 |
| 0.2324 | 12.0 | 1500 | 0.4432 | 0.3814 |
| 0.1338 | 16.0 | 2000 | 0.4169 | 0.3685 |
| 0.0893 | 20.0 | 2500 | 0.4634 | 0.3632 |
| 0.0631 | 24.0 | 3000 | 0.4804 | 0.3541 |
| 0.0497 | 28.0 | 3500 | 0.4634 | 0.3469 |
Framework versions
- Transformers 4.9.1
- Pytorch 1.9.0+cu102
- Datasets 1.17.0
- Tokenizers 0.10.3