- Notifications
You must be signed in to change notification settings - Fork4
mazumder-lab/CHITA
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
This is the offical repo of the ICML 2023 paperFast as CHITA: Neural Network Pruning with Combinatorial Optimization
This code has been tested with Python 3.7 and the following packages:
numba==0.56.4numpy==1.21.6scikit_learn==1.0.2torch==1.12.1+cu113torchvision==0.13.1+cu113We provide checkpoints for our best pruned models, obtained with the gradual pruning procedure described in the paper.
| Sparsity | Checkpoint |
|---|---|
| 75.28 | link |
| 89.00 | link |
| Sparsity | Checkpoint |
|---|---|
| 90.00 | link |
| 95.00 | link |
| 98.00 | link |
Scripts to run the algorithms are located inscripts/. The current code supports the following architectures (datasets): MLPNet (MNIST), ResNet20 (Cifar10), MobileNetV1 (Imagenet) and ResNet50 (Imagenet). Adding new models can be done throughmodel_factory function inutils/main_utils.py.
If you find CHITA useful in your research, please consider citing the following paper.
@InProceedings{pmlr-v202-benbaki23a, title = {Fast as {CHITA}: Neural Network Pruning with Combinatorial Optimization}, author = {Benbaki, Riade and Chen, Wenyu and Meng, Xiang and Hazimeh, Hussein and Ponomareva, Natalia and Zhao, Zhe and Mazumder, Rahul}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {2031--2049}, year = {2023},}