{"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/progressive-dnn-compression-a-key-to-achieve","title":"Progressive DNN Compression: A Key to Achieve Ultra-High Weight Pruning and Quantization Rates using ADMM","arxiv_id":"1903.09769","date":"2019-03-23","proceeding":null,"authors":["Shaokai Ye","Xiaoyu Feng","Tianyun Zhang","Xiaolong Ma","Sheng Lin","Zhengang Li","Kaidi Xu","Wujie Wen","Sijia Liu","Jian Tang","Makan Fardad","Xue Lin","Yongpan Liu","Yanzhi Wang"],"abstract":"Weight pruning and weight quantization are two important categories of DNN\nmodel compression. Prior work on these techniques are mainly based on\nheuristics. A recent work developed a systematic frame-work of DNN weight\npruning using the advanced optimization technique ADMM (Alternating Direction\nMethods of Multipliers), achieving one of state-of-art in weight pruning\nresults. In this work, we first extend such one-shot ADMM-based framework to\nguarantee solution feasibility and provide fast convergence rate, and\ngeneralize to weight quantization as well. We have further developed a\nmulti-step, progressive DNN weight pruning and quantization framework, with\ndual benefits of (i) achieving further weight pruning/quantization thanks to\nthe special property of ADMM regularization, and (ii) reducing the search space\nwithin each step. Extensive experimental results demonstrate the superior\nperformance compared with prior work. Some highlights: (i) we achieve 246x,36x,\nand 8x weight pruning on LeNet-5, AlexNet, and ResNet-50 models, respectively,\nwith (almost) zero accuracy loss; (ii) even a significant 61x weight pruning in\nAlexNet (ImageNet) results in only minor degradation in actual accuracy\ncompared with prior work; (iii) we are among the first to derive notable weight\npruning results for ResNet and MobileNet models; (iv) we derive the first\nlossless, fully binarized (for all layers) LeNet-5 for MNIST and VGG-16 for\nCIFAR-10; and (v) we derive the first fully binarized (for all layers) ResNet\nfor ImageNet with reasonable accuracy loss.","url_abs":"http://arxiv.org/abs/1903.09769v2","url_pdf":"http://arxiv.org/pdf/1903.09769v2.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":"progressive-dnn-compression-a-key-to-achieve","repo_url":"https://github.com/JiahaoYan98/Pruning-ADMM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"progressive-dnn-compression-a-key-to-achieve","repo_url":"https://github.com/yeshaokai/Robustness-Aware-Pruning-ADMM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"model-compression","task_name":"Model Compression"},{"task_slug":"quantization","task_name":"Quantization"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"admm","method_name":"ADMM"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"grouped-convolution","method_name":"Grouped Convolution"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"local-response-normalization","method_name":"Local Response Normalization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"pruning","method_name":"Pruning"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1903.09769","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}