Bayesian Optimization with Cluster and Rollback for CNN Pruning

Open source code for our ECCV2022 accepted paper. https://link.springer.com/chapter/10.1007/978-3-031-20050-2_29

Use BOCR:

First, softlink imagenet2012 dataset to ./dataset:

ln -s PATH_to_your_local_imagenet ./dataset/imagenet

Next, check the environment requirements in requirements.txt

Finally, run the script in the base folder:

For mobilenetv1 experiments:

./mobilenetv1.sh [options]

For mobilenetv2 experiments:

./mobilenetv2.sh [options]

For resnet56 experiments:

./resnet56.sh [options]

Options include non, static, db and gb, each stands for naive BO, layer clustering, direct rollback and gradual rollback.