metadata
license: apache-2.0
tags:
- generated_from_trainer
model-index:
- name: addison6
results: []
addison6
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: 1.5365
- Wer: 0.5065
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: 8
- 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 |
|---|---|---|---|---|
| 4.4213 | 3.5 | 500 | 2.9159 | 1.0 |
| 1.7942 | 6.99 | 1000 | 1.3946 | 0.7460 |
| 0.9148 | 10.49 | 1500 | 1.3088 | 0.6273 |
| 0.6178 | 13.99 | 2000 | 1.3045 | 0.5690 |
| 0.4443 | 17.48 | 2500 | 1.3874 | 0.5437 |
| 0.3461 | 20.98 | 3000 | 1.4667 | 0.5427 |
| 0.253 | 24.48 | 3500 | 1.5080 | 0.5189 |
| 0.2103 | 27.97 | 4000 | 1.5365 | 0.5065 |
Framework versions
- Transformers 4.17.0
- Pytorch 2.0.0+cu118
- Datasets 1.18.3
- Tokenizers 0.13.2