{"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/unveiling-redundancy-in-diffusion","title":"Unveiling Redundancy in Diffusion Transformers (DiTs): A Systematic Study","arxiv_id":"2411.13588","date":"2024-11-18","proceeding":null,"authors":["Xibo Sun","Jiarui Fang","Aoyu Li","Jinzhe Pan"],"abstract":"The increased model capacity of Diffusion Transformers (DiTs) and the demand for generating higher resolutions of images and videos have led to a significant rise in inference latency, impacting real-time performance adversely. While prior research has highlighted the presence of high similarity in activation values between adjacent diffusion steps (referred to as redundancy) and proposed various caching mechanisms to mitigate computational overhead, the exploration of redundancy in existing literature remains limited, with findings often not generalizable across different DiT models. This study aims to address this gap by conducting a comprehensive investigation into redundancy across a broad spectrum of mainstream DiT models. Our experimental analysis reveals substantial variations in the distribution of redundancy across diffusion steps among different DiT models. Interestingly, within a single model, the redundancy distribution remains stable regardless of variations in input prompts, step counts, or scheduling strategies. Given the lack of a consistent pattern across diverse models, caching strategies designed for a specific group of models may not easily transfer to others. To overcome this challenge, we introduce a tool for analyzing the redundancy of individual models, enabling subsequent research to develop tailored caching strategies for specific model architectures. The project is publicly available at https://github.com/xdit-project/DiTCacheAnalysis.","url_abs":"https://arxiv.org/abs/2411.13588v1","url_pdf":"https://arxiv.org/pdf/2411.13588v1.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":"unveiling-redundancy-in-diffusion","repo_url":"https://github.com/xdit-project/ditcacheanalysis","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"unveiling-redundancy-in-diffusion","repo_url":"https://github.com/xdit-project/xdit","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"scheduling","task_name":"Scheduling"}],"methods":[{"method_slug":"diffusion","method_name":"Diffusion"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2411.13588","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2411.13588"}},"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/xdit-project/ditcacheanalysis","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/xdit-project/xdit","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"unverified":8},"by_repo_kind":{"listed":{"samples":8,"ran":0,"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":"7875bbc064147a85","entry":"block_mask_to_delta_lut","repo":"xdit-project/xdit","repo_kind":"listed","path":"xfuser/core/vsa_attention.py","file_url":"https://github.com/xdit-project/xdit/blob/HEAD/xfuser/core/vsa_attention.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"7875bbc064147a85"}},{"code_sha256_prefix":"6048d66c2a6264ce","entry":"build_h3_vsa_metadata","repo":"xdit-project/xdit","repo_kind":"listed","path":"xfuser/core/vsa_h3_attention.py","file_url":"https://github.com/xdit-project/xdit/blob/HEAD/xfuser/core/vsa_h3_attention.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"6048d66c2a6264ce"}},{"code_sha256_prefix":"2ea4980332dcbb85","entry":"build_jenga_block_mask","repo":"xdit-project/xdit","repo_kind":"listed","path":"xfuser/core/vsa_attention.py","file_url":"https://github.com/xdit-project/xdit/blob/HEAD/xfuser/core/vsa_attention.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"2ea4980332dcbb85"}},{"code_sha256_prefix":"96d12ff78f7d783a","entry":"compute_h3_vsa_topk","repo":"xdit-project/xdit","repo_kind":"listed","path":"xfuser/core/vsa_h3_attention.py","file_url":"https://github.com/xdit-project/xdit/blob/HEAD/xfuser/core/vsa_h3_attention.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"96d12ff78f7d783a"}},{"code_sha256_prefix":"b9f11de480b9aae2","entry":"get_device","repo":"xdit-project/xdit","repo_kind":"listed","path":"xfuser/envs.py","file_url":"https://github.com/xdit-project/xdit/blob/HEAD/xfuser/envs.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"b9f11de480b9aae2"}},{"code_sha256_prefix":"5569fde490450b94","entry":"is_diffusers_import_error","repo":"xdit-project/xdit","repo_kind":"listed","path":"xfuser/compat.py","file_url":"https://github.com/xdit-project/xdit/blob/HEAD/xfuser/compat.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"5569fde490450b94"}},{"code_sha256_prefix":"bb417397aecb0c36","entry":"jenga_scheduled_drop_rate","repo":"xdit-project/xdit","repo_kind":"listed","path":"xfuser/core/vsa_attention.py","file_url":"https://github.com/xdit-project/xdit/blob/HEAD/xfuser/core/vsa_attention.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"bb417397aecb0c36"}},{"code_sha256_prefix":"a1f0fe3577596e03","entry":"tile_h3_vsa_tensor","repo":"xdit-project/xdit","repo_kind":"listed","path":"xfuser/core/vsa_h3_attention.py","file_url":"https://github.com/xdit-project/xdit/blob/HEAD/xfuser/core/vsa_h3_attention.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"a1f0fe3577596e03"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}