{"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/light-r1-curriculum-sft-dpo-and-rl-for-long","title":"Light-R1: Curriculum SFT, DPO and RL for Long COT from Scratch and Beyond","arxiv_id":"2503.10460","date":"2025-03-13","proceeding":null,"authors":["Liang Wen","Yunke Cai","Fenrui Xiao","Xin He","Qi An","Zhenyu Duan","Yimin Du","Junchen Liu","Lifu Tang","Xiaowei Lv","Haosheng Zou","Yongchao Deng","Shousheng Jia","Xiangzheng Zhang"],"abstract":"This paper introduces Light-R1, an open-source suite for training long reasoning models using reproducible and cost-effective methodology. Given the proprietary nature of data used in the DeepSeek-R1 series, we develop an alternative approach leveraging exclusively public data and models. Our curriculum training progressively increases data difficulty, combined with multi-staged post-training. Our Light-R1-32B model, trained from Qwen2.5-32B-Instruct, outperforms DeepSeek-R1-Distill-Qwen-32B in math reasoning. Experimental results show that this curriculum approach becomes more effective when distinct, diverse datasets are available for different training stages: fine-tuning DeepSeek-R1-Distilled models (pre-tuned by DeepSeek team on proprietary data) with 3,000 challenging examples from our curriculum dataset yielded state-of-the-art 7B and 14B models, while the 32B model, Light-R1-32B-DS performed comparably to QwQ-32B and DeepSeek-R1. Furthermore, we extend our work by applying GRPO on long reasoning models. Our final Light-R1-14B-DS achieves SOTA performance among 14B models in math, with AIME24 \\& 25 scores of 74.0 and 60.2 respectively, surpassing many 32B models and DeepSeek-R1-Distill-Llama-70B. Despite math-focused training, Light-R1-14B-DS demonstrates strong cross-domain generalization. Light-R1 represents a significant advancement in making sophisticated reasoning models more accessible and implementable in real-world applications. Our models, training data and code have been made available at https://github.com/Qihoo360/Light-R1.","url_abs":"https://arxiv.org/abs/2503.10460v3","url_pdf":"https://arxiv.org/pdf/2503.10460v3.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":"light-r1-curriculum-sft-dpo-and-rl-for-long","repo_url":"https://github.com/qihoo360/light-r1","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"domain-generalization","task_name":"Domain Generalization"},{"task_slug":"math","task_name":"Math"}],"methods":[{"method_slug":"dpo","method_name":"DPO"},{"method_slug":"focus","method_name":"Focus"},{"method_slug":"sft","method_name":"SFT"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2503.10460","atlas_url":"https://app.syntology.ai/?focus=2503.10460","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2503.10460"}},"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/qihoo360/light-r1","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"deterministic:regex_extraction","url":"https://github.com/Qihoo360/Light-R1","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"ran":5,"ran_honours":1,"ran_violates":1,"ran_fixture":1,"ran_draft_wrong":1,"unverified":4},"by_repo_kind":{"official":{"samples":13,"ran":9,"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":"57f0335f22430956","entry":"build_ngram_lookup","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"decontaminate/n_gram_check.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/decontaminate/n_gram_check.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"57f0335f22430956"}},{"code_sha256_prefix":"61619ba901c7b574","entry":"clean_text","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"decontaminate/character_matching.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/decontaminate/character_matching.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"61619ba901c7b574"}},{"code_sha256_prefix":"368d3db6347269c3","entry":"count_unknown_letters_in_expr","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/deepscaler/rewards/math_utils/utils.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/deepscaler/rewards/math_utils/utils.py","link_basis":"harvester_set","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"368d3db6347269c3"}},{"code_sha256_prefix":"360c671d6041ba3b","entry":"create_huggingface_actor","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/verl/verl/utils/model.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/verl/verl/utils/model.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"360c671d6041ba3b"}},{"code_sha256_prefix":"927fae1304011f13","entry":"normalize_string","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"decontaminate/n_gram_check.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/decontaminate/n_gram_check.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"927fae1304011f13"}},{"code_sha256_prefix":"95f02454eddf082a","entry":"should_allow_eval","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/deepscaler/rewards/math_utils/utils.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/deepscaler/rewards/math_utils/utils.py","link_basis":"harvester_set","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"95f02454eddf082a"}},{"code_sha256_prefix":"21331a58f93375e4","entry":"union_tensor_dict","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/verl/verl/protocol.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/verl/verl/protocol.py","link_basis":"harvester_set","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"DEP_MISSING","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"21331a58f93375e4"}},{"code_sha256_prefix":"25f0ea3f460f6ce1","entry":"unpad_dataproto","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/verl/verl/protocol.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/verl/verl/protocol.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"25f0ea3f460f6ce1"}},{"code_sha256_prefix":"b740be8e99a575cb","entry":"word_ngrams","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"decontaminate/n_gram_check.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/decontaminate/n_gram_check.py","link_basis":"harvester_set","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"b740be8e99a575cb"}},{"code_sha256_prefix":"19472d4343d0a082","entry":"get_huggingface_actor_config","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/verl/verl/utils/model.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/verl/verl/utils/model.py","link_basis":"harvester_set","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":"19472d4343d0a082"}},{"code_sha256_prefix":"919cf310ad4c189a","entry":"get_weight_loader","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/verl/verl/models/weight_loader_registry.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/verl/verl/models/weight_loader_registry.py","link_basis":"harvester_set","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":"919cf310ad4c189a"}},{"code_sha256_prefix":"ec578739f88adc22","entry":"mathd_normalize_answer","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/deepscaler/rewards/math_utils/utils.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/deepscaler/rewards/math_utils/utils.py","link_basis":"harvester_set","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":"ec578739f88adc22"}},{"code_sha256_prefix":"3b15e2ac7497c441","entry":"squeeze","repo":"Qihoo360/Light-R1","repo_kind":"official","path":"deepscaler-release/verl/verl/utils/model.py","file_url":"https://github.com/Qihoo360/Light-R1/blob/HEAD/deepscaler-release/verl/verl/utils/model.py","link_basis":"harvester_set","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":"3b15e2ac7497c441"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}