{"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/distillm-towards-streamlined-distillation-for","title":"DistiLLM: Towards Streamlined Distillation for Large Language Models","arxiv_id":"2402.03898","date":"2024-02-06","proceeding":null,"authors":["Jongwoo Ko","Sungnyun Kim","Tianyi Chen","Se-Young Yun"],"abstract":"Knowledge distillation (KD) is widely used for compressing a teacher model to a smaller student model, reducing its inference cost and memory footprint while preserving model capabilities. However, current KD methods for auto-regressive sequence models (e.g., large language models) suffer from missing a standardized objective function. Moreover, the recent use of student-generated outputs to address training-inference mismatches has significantly escalated computational costs. To tackle these issues, we introduce DistiLLM, a more effective and efficient KD framework for auto-regressive language models. DistiLLM comprises two components: (1) a novel skew Kullback-Leibler divergence loss, where we unveil and leverage its theoretical properties, and (2) an adaptive off-policy approach designed to enhance the efficiency in utilizing student-generated outputs. Extensive experiments, including instruction-following tasks, demonstrate the effectiveness of DistiLLM in building high-performing student models while achieving up to 4.3$\\times$ speedup compared to recent KD methods.","url_abs":"https://arxiv.org/abs/2402.03898v2","url_pdf":"https://arxiv.org/pdf/2402.03898v2.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":"distillm-towards-streamlined-distillation-for","repo_url":"https://github.com/jongwooko/distillm","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"jax","reach":null},{"paper_slug":"distillm-towards-streamlined-distillation-for","repo_url":"https://github.com/ghwang-s/abkd","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"distillm-towards-streamlined-distillation-for","repo_url":"https://github.com/jongwooko/distillm-2","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"distillm-towards-streamlined-distillation-for","repo_url":"https://github.com/mingyukim87/synergynerf","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"distillm-towards-streamlined-distillation-for","repo_url":"https://github.com/tianyic/only_train_once","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"instruction-following","task_name":"Instruction Following"},{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2402.03898","atlas_url":"https://app.syntology.ai/?focus=2402.03898","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2402.03898"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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/tianyic/only_train_once","reach":null},{"provenance":"deterministic:regex_extraction","url":"https://github.com/microsoft/LMOps","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/ghwang-s/abkd","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/mingyukim87/synergynerf","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/jongwooko/distillm","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/jongwooko/distillm-2","reach":{"status":"ok"}}],"summary":{"ran_fixture":1,"ran_draft_wrong":4},"by_repo_kind":{"official":{"samples":3,"ran":3,"repositories":1},"listed":{"samples":1,"ran":1,"repositories":1},"found_in_text":{"samples":1,"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":4,"samples":[{"code_sha256_prefix":"3595a0bc0ec56f02","entry":"ab_div","repo":"ghwang-s/abkd","repo_kind":"listed","path":"distillation_llm/distillm/losses.py","file_url":"https://github.com/ghwang-s/abkd/blob/HEAD/distillation_llm/distillm/losses.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3595a0bc0ec56f02"}},{"code_sha256_prefix":"575e937196d69d18","entry":"forward_kl","repo":"jongwooko/distillm","repo_kind":"official","path":"distillm/losses.py","file_url":"https://github.com/jongwooko/distillm/blob/HEAD/distillm/losses.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"575e937196d69d18"}},{"code_sha256_prefix":"cd958b79a0882e30","entry":"get_distil_loss","repo":"microsoft/LMOps","repo_kind":"found_in_text","path":"minillm/finetune.py","file_url":"https://github.com/microsoft/LMOps/blob/HEAD/minillm/finetune.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"cd958b79a0882e30"}},{"code_sha256_prefix":"283147d481c8695c","entry":"reverse_kl","repo":"jongwooko/distillm","repo_kind":"official","path":"distillm/losses.py","file_url":"https://github.com/jongwooko/distillm/blob/HEAD/distillm/losses.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"283147d481c8695c"}},{"code_sha256_prefix":"49aa907f43ee85d3","entry":"symmetric_kl","repo":"jongwooko/distillm","repo_kind":"official","path":"distillm/losses.py","file_url":"https://github.com/jongwooko/distillm/blob/HEAD/distillm/losses.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"49aa907f43ee85d3"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}