{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/layer-compensated-pruning-for-resource","title":"Layer-compensated Pruning for Resource-constrained Convolutional Neural Networks","arxiv_id":"1810.00518","date":"2018-10-01","proceeding":null,"authors":["Ting-Wu Chin","Cha Zhang","Diana Marculescu"],"abstract":"Resource-efficient convolution neural networks enable not only the\nintelligence on edge devices but also opportunities in system-level\noptimization such as scheduling. In this work, we aim to improve the\nperformance of resource-constrained filter pruning by merging two sub-problems\ncommonly considered, i.e., (i) how many filters to prune for each layer and\n(ii) which filters to prune given a per-layer pruning budget, into a global\nfilter ranking problem. Our framework entails a novel algorithm, dubbed\nlayer-compensated pruning, where meta-learning is involved to determine better\nsolutions. We show empirically that the proposed algorithm is superior to prior\nart in both effectiveness and efficiency. Specifically, we reduce the accuracy\ngap between the pruned and original networks from 0.9% to 0.7% with 8x\nreduction in time needed for meta-learning, i.e., from 1 hour down to 7\nminutes. To this end, we demonstrate the effectiveness of our algorithm using\nResNet and MobileNetV2 networks under CIFAR-10, ImageNet, and Bird-200\ndatasets.","url_abs":"http://arxiv.org/abs/1810.00518v2","url_pdf":"http://arxiv.org/pdf/1810.00518v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"layer-compensated-pruning-for-resource","repo_url":"https://github.com/cmu-enyac/LeGR","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"meta-learning","task_name":"Meta-Learning"},{"task_slug":"scheduling","task_name":"Scheduling"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"depthwise-convolution","method_name":"Depthwise Convolution"},{"method_slug":"depthwise-separable-convolution","method_name":"Depthwise Separable Convolution"},{"method_slug":"inverted-residual-block","method_name":"Inverted Residual Block"},{"method_slug":"pointwise-convolution","method_name":"Pointwise Convolution"},{"method_slug":"pruning","method_name":"Pruning"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1810.00518","atlas_url":"https://app.syntology.ai/?focus=1810.00518","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}