wav2vec2_s-f-o_8batch_5sec_0.0001lr_unfrozen
This model is a fine-tuned version of facebook/wav2vec2-large on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.4943
- Accuracy: 0.6667
- F1: 0.6742
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
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.003
- num_epochs: 1000
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
---|---|---|---|---|---|
2.1376 | 1.0 | 131 | 2.1461 | 0.15 | 0.0802 |
1.3836 | 2.0 | 262 | 1.8662 | 0.4 | 0.3888 |
1.2382 | 2.99 | 393 | 1.6891 | 0.45 | 0.4245 |
0.8998 | 4.0 | 525 | 1.4406 | 0.6 | 0.5890 |
0.5064 | 5.0 | 656 | 1.2466 | 0.7 | 0.6632 |
0.5248 | 6.0 | 787 | 1.1712 | 0.7 | 0.6705 |
0.5376 | 6.99 | 918 | 1.3778 | 0.6667 | 0.6620 |
0.4291 | 8.0 | 1050 | 2.0535 | 0.6167 | 0.5799 |
0.4947 | 9.0 | 1181 | 1.3218 | 0.7333 | 0.7250 |
0.5743 | 10.0 | 1312 | 1.7264 | 0.6667 | 0.6534 |
0.3847 | 10.99 | 1443 | 1.9041 | 0.6333 | 0.6319 |
0.6198 | 12.0 | 1575 | 1.3526 | 0.7167 | 0.6856 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3
- Downloads last month
- 6
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.