audio_cls_unispeech-sat-base-100h-libri-ft_minds14_finetune
This model is a fine-tuned version of microsoft/unispeech-sat-base-100h-libri-ft on the minds14 dataset. It achieves the following results on the evaluation set:
- Loss: 2.6384
- Accuracy: 0.0708
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: 3e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| No log | 1.0 | 2 | 2.6385 | 0.0442 |
| No log | 2.0 | 4 | 2.6385 | 0.0796 |
| No log | 3.0 | 6 | 2.6386 | 0.0619 |
| No log | 4.0 | 8 | 2.6375 | 0.0531 |
| 2.64 | 5.0 | 10 | 2.6372 | 0.0619 |
| 2.64 | 6.0 | 12 | 2.6376 | 0.0708 |
| 2.64 | 7.0 | 14 | 2.6379 | 0.0708 |
| 2.64 | 8.0 | 16 | 2.6381 | 0.0708 |
| 2.64 | 9.0 | 18 | 2.6384 | 0.0708 |
| 2.638 | 10.0 | 20 | 2.6384 | 0.0708 |
Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu117
- Datasets 2.12.0
- Tokenizers 0.13.3