{"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/deep-compression-compressing-deep-neural","title":"Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding","arxiv_id":"1510.00149","date":"2015-10-01","proceeding":null,"authors":["Song Han","Huizi Mao","William J. Dally"],"abstract":"Neural networks are both computationally intensive and memory intensive,\nmaking them difficult to deploy on embedded systems with limited hardware\nresources. To address this limitation, we introduce \"deep compression\", a three\nstage pipeline: pruning, trained quantization and Huffman coding, that work\ntogether to reduce the storage requirement of neural networks by 35x to 49x\nwithout affecting their accuracy. Our method first prunes the network by\nlearning only the important connections. Next, we quantize the weights to\nenforce weight sharing, finally, we apply Huffman coding. After the first two\nsteps we retrain the network to fine tune the remaining connections and the\nquantized centroids. Pruning, reduces the number of connections by 9x to 13x;\nQuantization then reduces the number of bits that represent each connection\nfrom 32 to 5. On the ImageNet dataset, our method reduced the storage required\nby AlexNet by 35x, from 240MB to 6.9MB, without loss of accuracy. Our method\nreduced the size of VGG-16 by 49x from 552MB to 11.3MB, again with no loss of\naccuracy. This allows fitting the model into on-chip SRAM cache rather than\noff-chip DRAM memory. Our compression method also facilitates the use of\ncomplex neural networks in mobile applications where application size and\ndownload bandwidth are constrained. Benchmarked on CPU, GPU and mobile GPU,\ncompressed network has 3x to 4x layerwise speedup and 3x to 7x better energy\nefficiency.","url_abs":"http://arxiv.org/abs/1510.00149v5","url_pdf":"http://arxiv.org/pdf/1510.00149v5.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":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/Irtza/Keras_model_compression","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/KarenUllrich/Tutorial_BayesianCompressionForDL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/Simon-lsy/Deep_Compression","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/bemova/Deep-Compression-Compressing-Deep-Neural-Networks-with-Pruning-Trained-Quantization-and-Huffman","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/cambridge-mlg/miracle","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/cambridge-mlg/variational-shannon-coding","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/ciodar/deep-compression","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/heguixiang/caffe_deep_compression","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/isha-garg/Deep_Compression","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/jiali-ms/JLM","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/lovepan1/caffe_ssd_traffic","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/may0324/DeepCompression-caffe","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/songhan/Deep-Compression-AlexNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"caffe2","reach":{"status":"ok","spdx":"BSD-2-Clause"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/songhan/SqueezeNet-Deep-Compression","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"caffe2","reach":{"status":"unanswered"}},{"paper_slug":"deep-compression-compressing-deep-neural","repo_url":"https://github.com/NervanaSystems/distiller","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"gone","observed_at":"2026-09-17","how":"tree_404+repo_404"}}],"tasks":[{"task_slug":null,"task_name":"CPU"},{"task_slug":null,"task_name":"GPU"},{"task_slug":"network-pruning","task_name":"Network Pruning"},{"task_slug":"quantization","task_name":"Quantization"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"grouped-convolution","method_name":"Grouped Convolution"},{"method_slug":"local-response-normalization","method_name":"Local Response Normalization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1510.00149","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1510.00149"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/songhan/SqueezeNet-Deep-Compression","reach":{"status":"unanswered"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/KarenUllrich/Tutorial_BayesianCompressionForDL","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/cambridge-mlg/miracle","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/may0324/DeepCompression-caffe","reach":{"status":"ok","spdx":"NOASSERTION"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/ciodar/deep-compression","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/bemova/Deep-Compression-Compressing-Deep-Neural-Networks-with-Pruning-Trained-Quantization-and-Huffman","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/isha-garg/Deep_Compression","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/jiali-ms/JLM","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/cambridge-mlg/variational-shannon-coding","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/lovepan1/caffe_ssd_traffic","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/NervanaSystems/distiller","reach":{"status":"gone","observed_at":"2026-09-17","how":"tree_404+repo_404"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Simon-lsy/Deep_Compression","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/heguixiang/caffe_deep_compression","reach":{"status":"ok","spdx":"NOASSERTION"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/songhan/Deep-Compression-AlexNet","reach":{"status":"ok","spdx":"BSD-2-Clause"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Irtza/Keras_model_compression","reach":{"status":"ok"}}],"summary":{"ran_honours":1,"ran_violates":1,"unverified":2},"by_repo_kind":{"listed":{"samples":4,"ran":2,"repositories":4}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":1,"samples":[{"code_sha256_prefix":"02b1e50dd5f50057","entry":"compress_matrix","repo":"KarenUllrich/Tutorial_BayesianCompressionForDL","repo_kind":"listed","path":"compression.py","file_url":"https://github.com/KarenUllrich/Tutorial_BayesianCompressionForDL/blob/HEAD/compression.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"02b1e50dd5f50057"}},{"code_sha256_prefix":"9df07be0f886f819","entry":"prune_weights","repo":"Simon-lsy/Deep_Compression","repo_kind":"listed","path":"compression.py","file_url":"https://github.com/Simon-lsy/Deep_Compression/blob/HEAD/compression.py","link_basis":"first_harvest_node","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"9df07be0f886f819"}},{"code_sha256_prefix":"8194a5a2ab394f86","entry":"DeepCompressor","repo":"bemova/Deep-Compression-Compressing-Deep-Neural-Networks-with-Pruning-Trained-Quantization-and-Huffman","repo_kind":"listed","path":"compressor.py","file_url":"https://github.com/bemova/Deep-Compression-Compressing-Deep-Neural-Networks-with-Pruning-Trained-Quantization-and-Huffman/blob/HEAD/compressor.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8194a5a2ab394f86"}},{"code_sha256_prefix":"21ce84076b9f1adc","entry":"kmeans_compress","repo":"jiali-ms/JLM","repo_kind":"listed","path":"train/comp.py","file_url":"https://github.com/jiali-ms/JLM/blob/HEAD/train/comp.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"21ce84076b9f1adc"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}