English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Machine learning pipelines for classification tasks often train a universalmodel to achieve accuracy across a broad range of classes. However, a typicaluser encounters only a limited selection of classes regularly. This disparityprovides an opportunity to enhance computational efficiency by tailoring modelsto focus on user-specific classes. Existing works rely on unstructured pruning,which introduces randomly distributed non-zero values in the model, making itunsuitable for hardware acceleration. Alternatively, some approaches employstructured pruning, such as channel pruning, but these tend to provide onlyminimal compression and may lead to reduced model accuracy. In this work, wepropose CRISP, a novel pruning framework leveraging a hybrid structuredsparsity pattern that combines both fine-grained N:M structured sparsity andcoarse-grained block sparsity. Our pruning strategy is guided by agradient-based class-aware saliency score, allowing us to retain weightscrucial for user-specific classes. CRISP achieves high accuracy with minimalmemory consumption for popular models like ResNet-50, VGG-16, and MobileNetV2on ImageNet and CIFAR-100 datasets. Moreover, CRISP delivers up to 14$\times$reduction in latency and energy consumption compared to existing pruningmethods while maintaining comparable accuracy. Our code is available athttps://github.com/shivmgg/CRISP/.

CRISP: Hybrid Structured Sparsity for Class-aware Model Pruning