Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2405.00074 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915349270626304 |
|---|---|
| author | Meng, Mark Huasong Guan, Hao Wan, Liuhuo Teo, Sin Gee Bai, Guangdong Dong, Jin Song |
| author_facet | Meng, Mark Huasong Guan, Hao Wan, Liuhuo Teo, Sin Gee Bai, Guangdong Dong, Jin Song |
| contents | We present PAODING, a toolkit to debloat pretrained neural network models through the lens of data-free pruning. To preserve the model fidelity, PAODING adopts an iterative process, which dynamically measures the effect of deleting a neuron to identify candidates that have the least impact to the output layer. Our evaluation shows that PAODING can significantly reduce the model size, generalize on different datasets and models, and meanwhile preserve the model fidelity in terms of test accuracy and adversarial robustness. PAODING is publicly available on PyPI via https://pypi.org/project/paoding-dl. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_00074 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | PAODING: A High-fidelity Data-free Pruning Toolkit for Debloating Pre-trained Neural Networks Meng, Mark Huasong Guan, Hao Wan, Liuhuo Teo, Sin Gee Bai, Guangdong Dong, Jin Song Machine Learning We present PAODING, a toolkit to debloat pretrained neural network models through the lens of data-free pruning. To preserve the model fidelity, PAODING adopts an iterative process, which dynamically measures the effect of deleting a neuron to identify candidates that have the least impact to the output layer. Our evaluation shows that PAODING can significantly reduce the model size, generalize on different datasets and models, and meanwhile preserve the model fidelity in terms of test accuracy and adversarial robustness. PAODING is publicly available on PyPI via https://pypi.org/project/paoding-dl. |
| title | PAODING: A High-fidelity Data-free Pruning Toolkit for Debloating Pre-trained Neural Networks |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2405.00074 |