Sparsifying Networks via Subdifferential Inclusion

Authors: Sagar Verma, Jean-Christophe Pesquet

ICML 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we conduct various experiments to validate the effectiveness of SIS in terms of test accuracy vs. sparsity and inference time FLOPs vs. sparsity by comparing against Rig L (Evci et al., 2020).
Researcher Affiliation Academia Sagar Verma 1 Jean-Christophe Pesquet 1 1Université Paris-Saclay, Centrale Supélec, Inria, Centre de Vision Numérique. Correspondence to: Sagar Verma <sagar.verma@centralesupelec.fr>.
Pseudocode Yes Algorithm 1 Douglas-Rachford algorithm for network compression
Open Source Code Yes Project page: https://sagarverma.github.io/compression
Open Datasets Yes We compare SIS with competitive baselines on CIFAR10/100 for three different sparsity regimes 90%, 95%, 98%, and the results are listed in Table 2.
Dataset Splits No The paper mentions using "test accuracy" and "20% samples per class were used during pruning phase" for some experiments, but does not explicitly provide the specific training/validation/test dataset splits (percentages or counts) required to fully reproduce the data partitioning for all experiments.
Hardware Specification No The paper mentions using "Graphical Processing Units (GPUs)" generally but does not provide specific hardware details like exact GPU or CPU models, processor types, or memory amounts used for running experiments.
Software Dependencies No The paper does not provide specific ancillary software details such as library or solver names with version numbers.
Experiment Setup Yes 20% samples per class were used during pruning phase of all the methods and were run for 40 epochs.