Dementia Wav2vec Scientific Specaugment V2
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: 1.7761 Accuracy: 0.5846 F1 Macro: 0.5842 F1 Weighted: 0.5848 Precision...
Source du modèle
- Type
- Other
- Taille du fichier
- 361 MB
- Modèle de base
- facebook/wav2vec2-base
- Version
- vRevision af7a127
- Mis à jour
- 17 oct. 2026
- Téléchargements
- 2
- Source
- Hugging Face
Description de la source
dementia-wav2vec-scientific-specaugment-V2
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: 1.7761
- Accuracy: 0.5846
- F1 Macro: 0.5842
- F1 Weighted: 0.5848
- Precision Macro: 0.5843
- Recall Macro: 0.5844
- F1 Nodementia: 0.5970
- F1 Dementia: 0.5714
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: 1e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 18
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Macro | F1 Weighted | Precision Macro | Recall Macro | F1 Nodementia | F1 Dementia |
|---|---|---|---|---|---|---|---|---|---|---|
| 0.6577 | 1.0 | 60 | 0.6785 | 0.6 | 0.5775 | 0.582 | 0.6093 | 0.5906 | 0.675 | 0.48 |
| 0.6029 | 2.0 | 120 | 0.6883 | 0.5538 | 0.5430 | 0.5463 | 0.5513 | 0.5479 | 0.6133 | 0.4727 |
| 0.7167 | 3.0 | 180 | 0.8361 | 0.4769 | 0.4708 | 0.4682 | 0.4817 | 0.4829 | 0.4138 | 0.5278 |
| 0.5111 | 4.0 | 240 | 0.8049 | 0.5385 | 0.5249 | 0.5286 | 0.5347 | 0.5318 | 0.6053 | 0.4444 |
| 0.5229 | 5.0 | 300 | 0.9100 | 0.5846 | 0.5309 | 0.5235 | 0.7123 | 0.6015 | 0.3721 | 0.6897 |
| 0.5605 | 6.0 | 360 | 1.0275 | 0.5538 | 0.5430 | 0.5463 | 0.5513 | 0.5479 | 0.6133 | 0.4727 |
| 0.3691 | 7.0 | 420 | 1.1598 | 0.5538 | 0.5534 | 0.5528 | 0.5570 | 0.5565 | 0.5397 | 0.5672 |
| 0.1476 | 8.0 | 480 | 1.2150 | 0.5692 | 0.5667 | 0.5682 | 0.5675 | 0.5669 | 0.6 | 0.5333 |
| 0.1799 | 9.0 | 540 | 1.2925 | 0.6154 | 0.6061 | 0.6033 | 0.6412 | 0.6238 | 0.5455 | 0.6667 |
| 0.2708 | 10.0 | 600 | 1.3388 | 0.6308 | 0.6199 | 0.6169 | 0.6639 | 0.6399 | 0.5556 | 0.6842 |
| 0.0982 | 11.0 | 660 | 1.5823 | 0.5692 | 0.5691 | 0.5694 | 0.5696 | 0.5697 | 0.5758 | 0.5625 |
| 0.1063 | 12.0 | 720 | 1.8316 | 0.5231 | 0.5157 | 0.5185 | 0.5192 | 0.5185 | 0.5753 | 0.4561 |
| 0.0269 | 13.0 | 780 | 1.5847 | 0.5846 | 0.5846 | 0.5846 | 0.5859 | 0.5859 | 0.5846 | 0.5846 |
| 0.0186 | 14.0 | 840 | 1.5050 | 0.6462 | 0.6458 | 0.6453 | 0.6504 | 0.6490 | 0.6349 | 0.6567 |
| 0.1148 | 15.0 | 900 | 1.5774 | 0.6154 | 0.6154 | 0.6154 | 0.6167 | 0.6167 | 0.6154 | 0.6154 |
| 0.0163 | 16.0 | 960 | 1.6449 | 0.6 | 0.5953 | 0.5933 | 0.6138 | 0.6063 | 0.5517 | 0.6389 |
| 0.0150 | 17.0 | 1020 | 1.7812 | 0.5692 | 0.5683 | 0.5692 | 0.5683 | 0.5683 | 0.5882 | 0.5484 |
| 0.0143 | 18.0 | 1080 | 1.7761 | 0.5846 | 0.5842 | 0.5848 | 0.5843 | 0.5844 | 0.5970 | 0.5714 |
Framework versions
- Transformers 5.5.4
- Pytorch 2.10.0+cu128
- Datasets 4.8.4
- Tokenizers 0.22.2
Sources
1 source
Sources
1 sourceVérifié 29 août
Artefacts du modèle
1 artefact
Artefacts du modèle
1 artefactExtraits de sources
2 extraits
Extraits de sources
2 extraits--- library_name: transformers license: apache-2.0 base_model: facebook/wav2vec2-base tags: - generated_from_trainer metrics: - accuracy model-index: - name: dementia-wav2vec-scientific-specaugment-V2 results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # dementia-wav2vec-scientific-specaugment-V2 This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset. It achieves the following results on the evaluation set: - Loss: 1.7761 - Accuracy: 0.5846 - F1 Macro: 0.5842 - F1 Weighted: 0.5848 - Precision Macro: 0.5843 - Recall Macro: 0.5844 - F1 Nodementia: 0.5970 - F1 Dementia: 0.5714 ## 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: 1e-05 - train_batch_size: 4 - eval_batch_size: 4 - seed: 42 - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments - lr_scheduler_type: linear - num_epochs: 18 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 Macro | F1 Weighted | Precision Macro | Recall Macro | F1 Nodementia | F1 Dementia | |:-------------:|:-----:|:----:|:---------------:|:--------:|:--------:|:-----------:|:---------------:|:------------:|:-------------:|:-----------:| | 0.6577 | 1.0 | 60 | 0.6785 | 0.6 | 0.5775 | 0.582 | 0.6093 | 0.5906 | 0.675 | 0.48 | | 0.6029 | 2.0 | 120 | 0.6883 | 0.5538 | 0.5430 | 0.5463 | 0.5513 | 0.5479 | 0.6133 | 0.4727 | | 0.7167 | 3.0 | 180 | 0.8361 | 0.4769 | 0.4708 | 0.4682 | 0.4817 | 0.4829 | 0.4138 | 0.5278 | | 0.5111 | 4.0 | 240 | 0.8049 | 0.5385 | 0.5249 | 0.5286 | 0.5347 | 0.5318 | 0.6053 | 0.4444 | | 0.5229 | 5.0 | 300 | 0.9100 | 0.5846 | 0.5309 | 0.5235 | 0.7123 | 0.6015 | 0.3721 | 0.6897 | | 0.5605 | 6.0 | 360 | 1.0275 | 0.5538 | 0.5430 | 0.5463 | 0.5513 | 0.5479 | 0.6133 | 0.4727 | | 0.3691 | 7.0 | 420 | 1.1598 | 0.5538 | 0.5534 | 0.5528 | 0.5570 | 0.5565 | 0.5397 | 0.5672 | | 0.1476 | 8.0 | 480 | 1.2150 | 0.5692 | 0.5667 | 0.5682 | 0.5675 |...
Source context: 2 downloads · 0 likes · Pipeline audio-classification · Library transformers · Repo Mrsmetamorphosis/dementia-wav2vec-scientific-specaugment-V2
