metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: results2
results: []
results2
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: 2.4488
- Wer: 0.9986
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: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 8.1902 | 2.5 | 500 | 8.5776 | 0.9999 |
| 3.3623 | 5.0 | 1000 | 3.0266 | 0.9981 |
| 2.6484 | 7.5 | 1500 | 2.6942 | 0.9981 |
| 2.4592 | 10.0 | 2000 | 2.5737 | 0.9978 |
| 2.3747 | 12.5 | 2500 | 2.5134 | 0.9985 |
| 2.3306 | 15.0 | 3000 | 2.4761 | 0.9986 |
| 2.3032 | 17.5 | 3500 | 2.4554 | 0.9985 |
| 2.2926 | 20.0 | 4000 | 2.4488 | 0.9986 |
Framework versions
- Transformers 4.37.0
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.15.2