{"url":"/task/network-pruning","name":"Network Pruning","slug":"network-pruning","description_markdown":"**Network Pruning** is a popular approach to reduce a heavy network to obtain a light-weight form by removing redundancy in the heavy network. In this approach, a complex over-parameterized network is first trained, then pruned based on come criterions, and finally fine-tuned to achieve comparable performance with reduced parameters.\n\n\n<span class=\"description-source\">Source: [Ensemble Knowledge Distillation for Learning Improved and Efficient Networks ](https://arxiv.org/abs/1909.08097)</span>","categories":[{"name":"Methodology","url":"/area/methodology"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":534,"papers_with_code":239,"benchmarks":5,"benchmark_tables_in_archive":5,"benchmark_tables_shown":5,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":5,"subtasks":0,"parent_tasks":0},"benchmarks":[{"leaderboard":"/sota/network-pruning-on-imagenet","slug":"network-pruning-on-imagenet","dataset":"ImageNet","dataset_url":"/dataset/imagenet","rows_in_archive":16,"metrics":["Accuracy","GFLOPs","MParams"],"first_row_in_archive_order":{"model":"ResNet50-2.3 GFLOPs","paper_title":"Pruning Filters for Efficient ConvNets","paper_url":"/paper/pruning-filters-for-efficient-convnets","paper_date":"2016-08-31","arxiv_id":"1608.08710","code_links":[{"title":"PaddlePaddle/PaddleOCR","url":"https://github.com/PaddlePaddle/PaddleOCR"},{"title":"VainF/Torch-Pruning","url":"https://github.com/VainF/Torch-Pruning"},{"title":"he-y/filter-pruning-geometric-median","url":"https://github.com/he-y/filter-pruning-geometric-median"},{"title":"midasklr/yolov5prune","url":"https://github.com/midasklr/yolov5prune"},{"title":"oandrienko/fast-semantic-segmentation","url":"https://github.com/oandrienko/fast-semantic-segmentation"},{"title":"marcoancona/TorchPruner","url":"https://github.com/marcoancona/TorchPruner"},{"title":"mingsun-tse/regularization-pruning","url":"https://github.com/mingsun-tse/regularization-pruning"},{"title":"Adlik/model_optimizer","url":"https://github.com/Adlik/model_optimizer"},{"title":"guoxiaolu/model_compression","url":"https://github.com/guoxiaolu/model_compression"},{"title":"matthew-mcateer/Keras_pruning","url":"https://github.com/matthew-mcateer/Keras_pruning"},{"title":"AlumLuther/PruningFilters","url":"https://github.com/AlumLuther/PruningFilters"},{"title":"EstherBear/implementation-of-pruning-filters","url":"https://github.com/EstherBear/implementation-of-pruning-filters"},{"title":"arturjordao/PruningNeuralNetworks","url":"https://github.com/arturjordao/PruningNeuralNetworks"},{"title":"lehduong/ginp","url":"https://github.com/lehduong/ginp"},{"title":"lehduong/kesi","url":"https://github.com/lehduong/kesi"},{"title":"siyuan0/pytorch_model_prune","url":"https://github.com/siyuan0/pytorch_model_prune"},{"title":"mvpzhangqiu/yolov5prune","url":"https://github.com/mvpzhangqiu/yolov5prune"},{"title":"cailinhang/2018-Graduation-Project","url":"https://github.com/cailinhang/2018-Graduation-Project"},{"title":"AnishDelft/ModelCompression","url":"https://github.com/AnishDelft/ModelCompression"},{"title":"mattangus/fast-semantic-segmentation","url":"https://github.com/mattangus/fast-semantic-segmentation"},{"title":"prerakmody/CS4180-DL","url":"https://github.com/prerakmody/CS4180-DL"}],"syntology":{"n":36,"n_ran":20,"n_unverified":16,"n_pointer_only":20}}},{"leaderboard":"/sota/network-pruning-on-imagenet-resnet-50-90","slug":"network-pruning-on-imagenet-resnet-50-90","dataset":"ImageNet - ResNet 50 - 90% sparsity","dataset_url":null,"rows_in_archive":9,"metrics":["Top-1 Accuracy"],"first_row_in_archive_order":{"model":"Feather","paper_title":"Feather: An Elegant Solution to Effective DNN Sparsification","paper_url":"/paper/feather-an-elegant-solution-to-effective-dnn","paper_date":"2023-10-03","arxiv_id":"2310.02448","code_links":[{"title":"athglentis/feather","url":"https://github.com/athglentis/feather"}],"syntology":null}},{"leaderboard":"/sota/network-pruning-on-cifar-100","slug":"network-pruning-on-cifar-100","dataset":"CIFAR-100","dataset_url":"/dataset/cifar-100","rows_in_archive":5,"metrics":["Accuracy","GFLOPs","Inference Time (ms)"],"first_row_in_archive_order":{"model":"Dense","paper_title":"AC/DC: Alternating Compressed/DeCompressed Training of Deep Neural Networks","paper_url":"/paper/ac-dc-alternating-compressed-decompressed","paper_date":"2021-06-23","arxiv_id":"2106.12379","code_links":[{"title":"IST-DASLab/ACDC","url":"https://github.com/IST-DASLab/ACDC"},{"title":"IST-DASLab/sparseprop","url":"https://github.com/IST-DASLab/sparseprop"}],"syntology":{"n":5,"n_ran":2,"n_unverified":3,"n_pointer_only":0}}},{"leaderboard":"/sota/network-pruning-on-cifar-10","slug":"network-pruning-on-cifar-10","dataset":"CIFAR-10","dataset_url":"/dataset/cifar-10","rows_in_archive":4,"metrics":["Accuracy","GFLOPs","Inference Time (ms)"],"first_row_in_archive_order":{"model":"TAS-pruned ResNet-110","paper_title":"Network Pruning via Transformable Architecture Search","paper_url":"/paper/network-pruning-via-transformable","paper_date":"2019-05-23","arxiv_id":"1905.09717","code_links":[{"title":"D-X-Y/GDAS","url":"https://github.com/D-X-Y/GDAS"},{"title":"D-X-Y/NAS-Projects","url":"https://github.com/D-X-Y/NAS-Projects"},{"title":"D-X-Y/AutoDL-Projects","url":"https://github.com/D-X-Y/AutoDL-Projects"},{"title":"xxlya/COS598D_Assignment1","url":"https://github.com/xxlya/COS598D_Assignment1"}],"syntology":null}},{"leaderboard":"/sota/network-pruning-on-mnist","slug":"network-pruning-on-mnist","dataset":"MNIST","dataset_url":"/dataset/mnist","rows_in_archive":1,"metrics":["Avg #Steps"],"first_row_in_archive_order":{"model":"FFN-ShapleyPruned","paper_title":"Analysing Neural Network Topologies: a Game Theoretic Approach","paper_url":"/paper/analysing-neural-network-topologies-a-game","paper_date":"2019-04-17","arxiv_id":"1904.08166","code_links":[],"syntology":null}}],"datasets":[{"url":"/dataset/cifar-10","name":"CIFAR-10","full_name":"CIFAR-10","num_papers_in_archive":16145},{"url":"/dataset/imagenet","name":"ImageNet","full_name":"","num_papers_in_archive":15430},{"url":"/dataset/cifar-100","name":"CIFAR-100","full_name":"","num_papers_in_archive":9045},{"url":"/dataset/mnist","name":"MNIST","full_name":"","num_papers_in_archive":7651},{"url":"/dataset/netzschleuder","name":"Netzschleuder","full_name":"network catalogue, repository and centrifuge","num_papers_in_archive":5}],"subtasks":[],"parent_tasks":[],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":30,"of":239,"tagged_in_all":534,"items":[{"url":"/paper/squeezenet-alexnet-level-accuracy-with-50x","title":"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size","date":"2016-02-24","arxiv_id":"1602.07360","repositories_listed":59,"syntology":{"n":4,"n_ran":4,"n_unverified":0,"n_pointer_only":2}},{"url":"/paper/the-lottery-ticket-hypothesis-finding-sparse","title":"The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks","date":"2018-03-09","arxiv_id":"1803.03635","repositories_listed":24,"syntology":{"n":25,"n_ran":4,"n_unverified":21,"n_pointer_only":6}},{"url":"/paper/pruning-filters-for-efficient-convnets","title":"Pruning Filters for Efficient ConvNets","date":"2016-08-31","arxiv_id":"1608.08710","repositories_listed":21,"syntology":{"n":36,"n_ran":20,"n_unverified":16,"n_pointer_only":20}},{"url":"/paper/deep-compression-compressing-deep-neural","title":"Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding","date":"2015-10-01","arxiv_id":"1510.00149","repositories_listed":15,"syntology":{"n":4,"n_ran":2,"n_unverified":2,"n_pointer_only":1}},{"url":"/paper/snip-single-shot-network-pruning-based-on","title":"SNIP: Single-shot Network Pruning based on Connection Sensitivity","date":"2018-10-04","arxiv_id":"1810.02340","repositories_listed":8,"syntology":{"n":12,"n_ran":5,"n_unverified":7,"n_pointer_only":7}},{"url":"/paper/a-simple-and-effective-pruning-approach-for","title":"A Simple and Effective Pruning Approach for Large Language Models","date":"2023-06-20","arxiv_id":"2306.11695","repositories_listed":7,"syntology":{"n":22,"n_ran":9,"n_unverified":13,"n_pointer_only":9}},{"url":"/paper/manifold-regularized-dynamic-network-pruning","title":"Manifold Regularized Dynamic Network Pruning","date":"2021-03-10","arxiv_id":"2103.05861","repositories_listed":7,"syntology":{"n":2,"n_ran":2,"n_unverified":0,"n_pointer_only":2}},{"url":"/paper/scop-scientific-control-for-reliable-neural","title":"SCOP: Scientific Control for Reliable Neural Network Pruning","date":"2020-10-21","arxiv_id":"2010.10732","repositories_listed":4,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":1}},{"url":"/paper/movement-pruning-adaptive-sparsity-by-fine","title":"Movement Pruning: Adaptive Sparsity by Fine-Tuning","date":"2020-05-15","arxiv_id":"2005.07683","repositories_listed":4,"syntology":{"n":4,"n_ran":4,"n_unverified":0,"n_pointer_only":4}},{"url":"/paper/similarity-of-neural-networks-with-gradients","title":"Similarity of Neural Networks with Gradients","date":"2020-03-25","arxiv_id":"2003.11498","repositories_listed":4,"syntology":{"n":11,"n_ran":0,"n_unverified":11,"n_pointer_only":0}},{"url":"/paper/on-pruning-adversarially-robust-neural","title":"HYDRA: Pruning Adversarially Robust Neural Networks","date":"2020-02-24","arxiv_id":"2002.10509","repositories_listed":4,"syntology":{"n":19,"n_ran":4,"n_unverified":15,"n_pointer_only":17}},{"url":"/paper/190600586","title":"Discovering Neural Wirings","date":"2019-06-03","arxiv_id":"1906.00586","repositories_listed":4,"syntology":{"n":3,"n_ran":3,"n_unverified":0,"n_pointer_only":3}},{"url":"/paper/network-pruning-via-transformable","title":"Network Pruning via Transformable Architecture Search","date":"2019-05-23","arxiv_id":"1905.09717","repositories_listed":4,"syntology":null},{"url":"/paper/a-systematic-dnn-weight-pruning-framework","title":"A Systematic DNN Weight Pruning Framework using Alternating Direction Method of Multipliers","date":"2018-04-10","arxiv_id":"1804.03294","repositories_listed":4,"syntology":{"n":3,"n_ran":3,"n_unverified":0,"n_pointer_only":3}},{"url":"/paper/packnet-adding-multiple-tasks-to-a-single","title":"PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning","date":"2017-11-15","arxiv_id":"1711.05769","repositories_listed":4,"syntology":{"n":1,"n_ran":1,"n_unverified":0,"n_pointer_only":1}},{"url":"/paper/picking-winning-tickets-before-training-by-1","title":"Picking Winning Tickets Before Training by Preserving Gradient Flow","date":"2020-02-18","arxiv_id":"2002.07376","repositories_listed":3,"syntology":{"n":2,"n_ran":1,"n_unverified":1,"n_pointer_only":0}},{"url":"/paper/importance-estimation-for-neural-network-1","title":"Importance Estimation for Neural Network Pruning","date":"2019-06-25","arxiv_id":"1906.10771","repositories_listed":3,"syntology":null},{"url":"/paper/adversarial-fine-tuning-of-compressed-neural","title":"Adversarial Fine-tuning of Compressed Neural Networks for Joint Improvement of Robustness and Efficiency","date":"2024-03-14","arxiv_id":"2403.09441","repositories_listed":2,"syntology":null},{"url":"/paper/falcon-flop-aware-combinatorial-optimization","title":"FALCON: FLOP-Aware Combinatorial Optimization for Neural Network Pruning","date":"2024-03-11","arxiv_id":"2403.07094","repositories_listed":2,"syntology":null},{"url":"/paper/do-localization-methods-actually-localize","title":"Do Localization Methods Actually Localize Memorized Data in LLMs? A Tale of Two Benchmarks","date":"2023-11-15","arxiv_id":"2311.09060","repositories_listed":2,"syntology":{"n":15,"n_ran":6,"n_unverified":9,"n_pointer_only":0}},{"url":"/paper/beyond-size-how-gradients-shape-pruning","title":"Beyond Size: How Gradients Shape Pruning Decisions in Large Language Models","date":"2023-11-08","arxiv_id":"2311.04902","repositories_listed":2,"syntology":{"n":7,"n_ran":3,"n_unverified":4,"n_pointer_only":0}},{"url":"/paper/improving-the-transferability-of-adversarial-5","title":"Improving the Transferability of Adversarial Examples via Direction Tuning","date":"2023-03-27","arxiv_id":"2303.15109","repositories_listed":2,"syntology":null},{"url":"/paper/iterative-soft-shrinkage-learning-for","title":"Iterative Soft Shrinkage Learning for Efficient Image Super-Resolution","date":"2023-03-16","arxiv_id":"2303.09650","repositories_listed":2,"syntology":null},{"url":"/paper/upop-unified-and-progressive-pruning-for","title":"UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers","date":"2023-01-31","arxiv_id":"2301.13741","repositories_listed":2,"syntology":{"n":2,"n_ran":0,"n_unverified":2,"n_pointer_only":0}},{"url":"/paper/why-is-the-state-of-neural-network-pruning-so","title":"Why is the State of Neural Network Pruning so Confusing? On the Fairness, Comparison Setup, and Trainability in Network Pruning","date":"2023-01-12","arxiv_id":"2301.05219","repositories_listed":2,"syntology":null},{"url":"/paper/training-deep-neural-networks-with-joint","title":"Training Deep Neural Networks with Joint Quantization and Pruning of Weights and Activations","date":"2021-10-15","arxiv_id":"2110.08271","repositories_listed":2,"syntology":null},{"url":"/paper/group-fisher-pruning-for-practical-network","title":"Group Fisher Pruning for Practical Network Compression","date":"2021-08-02","arxiv_id":"2108.00708","repositories_listed":2,"syntology":null},{"url":"/paper/manipulating-identical-filter-redundancy-for","title":"Manipulating Identical Filter Redundancy for Efficient Pruning on Deep and Complicated CNN","date":"2021-07-30","arxiv_id":"2107.14444","repositories_listed":2,"syntology":null},{"url":"/paper/efficient-matrix-free-approximations-of","title":"M-FAC: Efficient Matrix-Free Approximations of Second-Order Information","date":"2021-07-07","arxiv_id":"2107.03356","repositories_listed":2,"syntology":{"n":14,"n_ran":7,"n_unverified":7,"n_pointer_only":0}},{"url":"/paper/ac-dc-alternating-compressed-decompressed","title":"AC/DC: Alternating Compressed/DeCompressed Training of Deep Neural Networks","date":"2021-06-23","arxiv_id":"2106.12379","repositories_listed":2,"syntology":{"n":5,"n_ran":2,"n_unverified":3,"n_pointer_only":0}}],"syntology_records":20,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}