{"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/skyladder-better-and-faster-pretraining-via","title":"SkyLadder: Better and Faster Pretraining via Context Window Scheduling","arxiv_id":"2503.15450","date":"2025-03-19","proceeding":null,"authors":["Tongyao Zhu","Qian Liu","Haonan Wang","Shiqi Chen","Xiangming Gu","Tianyu Pang","Min-Yen Kan"],"abstract":"Recent advancements in LLM pretraining have featured ever-expanding context windows to process longer sequences. However, our pilot study reveals that models pretrained with shorter context windows consistently outperform their long-context counterparts under a fixed token budget. This finding motivates us to explore an optimal context window scheduling strategy to better balance long-context capability with pretraining efficiency. To this end, we propose SkyLadder, a simple yet effective approach that implements a short-to-long context window transition. SkyLadder preserves strong standard benchmark performance, while matching or exceeding baseline results on long context tasks. Through extensive experiments, we pre-train 1B-parameter models (up to 32K context) and 3B-parameter models (8K context) on 100B tokens, demonstrating that SkyLadder yields consistent gains of up to 3.7% on common benchmarks, while achieving up to 22% faster training speeds compared to baselines. The code is at https://github.com/sail-sg/SkyLadder.","url_abs":"https://arxiv.org/abs/2503.15450v1","url_pdf":"https://arxiv.org/pdf/2503.15450v1.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":"skyladder-better-and-faster-pretraining-via","repo_url":"https://github.com/sail-sg/skyladder","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":null,"task_name":"8k"},{"task_slug":"scheduling","task_name":"Scheduling"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2503.15450","atlas_url":"https://app.syntology.ai/?focus=2503.15450","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2503.15450"}},"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/sail-sg/skyladder","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"ran":5,"ran_draft_wrong":1,"ran_honours":1,"unverified":3},"by_repo_kind":{"official":{"samples":10,"ran":7,"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":"e05d85cc898c445c","entry":"adapter_filter","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/adapter_v2.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/adapter_v2.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":"e05d85cc898c445c"}},{"code_sha256_prefix":"5728c74084e12d35","entry":"apply_rope","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/model.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/model.py","link_basis":"harvester_set","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"MISDECLARED","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"5728c74084e12d35"}},{"code_sha256_prefix":"a30c7463cb6330d7","entry":"build_rope_cache","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/model.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/model.py","link_basis":"harvester_set","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"a30c7463cb6330d7"}},{"code_sha256_prefix":"fb2d39aee5a88506","entry":"get_fragment_lens_fixed_length","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/packed_dataset.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/packed_dataset.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":"fb2d39aee5a88506"}},{"code_sha256_prefix":"81715c046abc63f1","entry":"get_fragment_lens_optimized","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/packed_dataset.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/packed_dataset.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":"81715c046abc63f1"}},{"code_sha256_prefix":"efd8acd4de8ba2a8","entry":"lora_filter","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/lora.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/lora.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":"efd8acd4de8ba2a8"}},{"code_sha256_prefix":"c6283b7a1a1785ba","entry":"maybe_align","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/rmsnorm.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/rmsnorm.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":"c6283b7a1a1785ba"}},{"code_sha256_prefix":"38ffb0df89b0d01f","entry":"adapter_filter","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/adapter.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/adapter.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":"38ffb0df89b0d01f"}},{"code_sha256_prefix":"3f3d25ea807f6d5c","entry":"get_flops_available","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/speed_monitor.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/speed_monitor.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":"3f3d25ea807f6d5c"}},{"code_sha256_prefix":"3bf153c27392d5fd","entry":"get_fragment_lens_fixed_length_intramask","repo":"sail-sg/skyladder","repo_kind":"official","path":"lit_gpt/packed_dataset.py","file_url":"https://github.com/sail-sg/skyladder/blob/HEAD/lit_gpt/packed_dataset.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":"3bf153c27392d5fd"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}