large_base_M08
This model is a fine-tuned version of facebook/wav2vec2-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 2.0553
- Wer: 1.0
- Cer: 0.8260
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: 20
- 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: 17080
- num_epochs: 100
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 8.55 | 20.0 | 68320 | 2.2120 | 1.0 | 0.8433 |
| 2.7106 | 40.0 | 136640 | 2.1153 | 1.0 | 0.8329 |
| 2.6764 | 60.0 | 204960 | 2.0673 | 1.0 | 0.8294 |
| 2.664 | 80.0 | 273280 | 2.0591 | 1.0 | 0.8268 |
| 2.6581 | 100.0 | 341600 | 2.0553 | 1.0 | 0.8260 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.2+cu102
- Datasets 2.3.2
- Tokenizers 0.12.1