{"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/dynamic-sparse-training-with-structured","title":"Dynamic Sparse Training with Structured Sparsity","arxiv_id":"2305.02299","date":"2023-05-03","proceeding":null,"authors":["Mike Lasby","Anna Golubeva","Utku Evci","Mihai Nica","Yani Ioannou"],"abstract":"Dynamic Sparse Training (DST) methods achieve state-of-the-art results in sparse neural network training, matching the generalization of dense models while enabling sparse training and inference. Although the resulting models are highly sparse and theoretically less computationally expensive, achieving speedups with unstructured sparsity on real-world hardware is challenging. In this work, we propose a sparse-to-sparse DST method, Structured RigL (SRigL), to learn a variant of fine-grained structured N:M sparsity by imposing a constant fan-in constraint. Using our empirical analysis of existing DST methods at high sparsity, we additionally employ a neuron ablation method which enables SRigL to achieve state-of-the-art sparse-to-sparse structured DST performance on a variety of Neural Network (NN) architectures. Using a 90% sparse linear layer, we demonstrate a real-world acceleration of 3.4x/2.5x on CPU for online inference and 1.7x/13.0x on GPU for inference with a batch size of 256 when compared to equivalent dense/unstructured (CSR) sparse layers, respectively.","url_abs":"https://arxiv.org/abs/2305.02299v4","url_pdf":"https://arxiv.org/pdf/2305.02299v4.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":"dynamic-sparse-training-with-structured","repo_url":"https://github.com/calgaryml/condensed-sparsity","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":null,"task_name":"CPU"},{"task_slug":null,"task_name":"GPU"}],"methods":[{"method_slug":"dst","method_name":"DST"},{"method_slug":"rigl","method_name":"RigL"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2305.02299","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2305.02299"}},"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/calgaryml/condensed-sparsity","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran_honours":1,"unverified":9},"by_repo_kind":{"official":{"samples":10,"ran":1,"repositories":1}},"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":0,"samples":[{"code_sha256_prefix":"90fadec0aec03c1b","entry":"collate_fn","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"src/rigl_torch/datasets/_coco.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/src/rigl_torch/datasets/_coco.py","link_basis":"plan_row","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"90fadec0aec03c1b"}},{"code_sha256_prefix":"ffc9a12b2490e5ad","entry":"collate","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"src/rigl_torch/datasets/_coco_detection_v2.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/src/rigl_torch/datasets/_coco_detection_v2.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":"ffc9a12b2490e5ad"}},{"code_sha256_prefix":"f1cb6ce89b53ffe9","entry":"forward_sparsity_single","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"src/condensed_sparsity/condensed_linear.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/src/condensed_sparsity/condensed_linear.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":"f1cb6ce89b53ffe9"}},{"code_sha256_prefix":"a094e0de1437a438","entry":"forward_sparsity_v","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"src/condensed_sparsity/condensed_linear.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/src/condensed_sparsity/condensed_linear.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":"a094e0de1437a438"}},{"code_sha256_prefix":"3654bf74d0f7f041","entry":"get_conv_output_size","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"src/micronet_challenge/counting.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/src/micronet_challenge/counting.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":"3654bf74d0f7f041"}},{"code_sha256_prefix":"4b82fc73fa17da12","entry":"get_flops_per_activation","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"src/micronet_challenge/counting.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/src/micronet_challenge/counting.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":"4b82fc73fa17da12"}},{"code_sha256_prefix":"2bcd7c08957f4644","entry":"get_sparse_size","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"src/micronet_challenge/counting.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/src/micronet_challenge/counting.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":"2bcd7c08957f4644"}},{"code_sha256_prefix":"3cdc178d93c08dbe","entry":"set_seed","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"train_rigl.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/train_rigl.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":"3cdc178d93c08dbe"}},{"code_sha256_prefix":"6204181691c991bc","entry":"structured_condensed_conv2d_factory","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"src/condensed_sparsity/condensed_linear.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/src/condensed_sparsity/condensed_linear.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":"6204181691c991bc"}},{"code_sha256_prefix":"e62b6d8b58cd26a5","entry":"test","repo":"calgaryml/condensed-sparsity","repo_kind":"official","path":"train_rigl.py","file_url":"https://github.com/calgaryml/condensed-sparsity/blob/HEAD/train_rigl.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":"e62b6d8b58cd26a5"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}