{"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/mars-unleashing-the-power-of-variance","title":"MARS: Unleashing the Power of Variance Reduction for Training Large Models","arxiv_id":"2411.10438","date":"2024-11-15","proceeding":null,"authors":["Huizhuo Yuan","Yifeng Liu","Shuang Wu","Xun Zhou","Quanquan Gu"],"abstract":"Training deep neural networks--and more recently, large models demands efficient and scalable optimizers. Adaptive gradient algorithms like Adam, AdamW, and their variants have been central to this task. Despite the development of numerous variance reduction algorithms in the past decade aimed at accelerating stochastic optimization in both convex and nonconvex settings, variance reduction has not found widespread success in training deep neural networks or large language models. Consequently, it has remained a less favored approach in modern AI. In this paper, to unleash the power of variance reduction for efficient training of large models, we propose a unified optimization framework, MARS (Make vAriance Reduction Shine), which reconciles preconditioned gradient methods with variance reduction via a scaled stochastic recursive momentum technique. Within our framework, we introduce three instances of MARS that leverage preconditioned gradient updates based on AdamW, Lion, and Shampoo, respectively. We also draw a connection between our algorithms and existing optimizers. Experimental results on training GPT-2 models indicate that MARS consistently outperforms AdamW by a large margin. The implementation of MARS is available at https://github.com/AGI-Arena/MARS.","url_abs":"https://arxiv.org/abs/2411.10438v2","url_pdf":"https://arxiv.org/pdf/2411.10438v2.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":"mars-unleashing-the-power-of-variance","repo_url":"https://github.com/AGI-Arena/MARS","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"mars-unleashing-the-power-of-variance","repo_url":"https://github.com/huggingface/pytorch-image-models","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"stochastic-optimization","task_name":"Stochastic Optimization"}],"methods":[{"method_slug":"adam","method_name":"Adam"},{"method_slug":"adamw","method_name":"AdamW"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"attention-dropout","method_name":"Attention Dropout"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"cosine-annealing","method_name":"Cosine Annealing"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"discriminative-fine-tuning","method_name":"Discriminative Fine-Tuning"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"gpt-2","method_name":"GPT-2"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"linear-warmup-with-cosine-annealing","method_name":"Linear Warmup With Cosine Annealing"},{"method_slug":"lion","method_name":"Lion"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"weight-decay","method_name":"Weight Decay"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2411.10438","atlas_url":"https://app.syntology.ai/?focus=2411.10438","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2411.10438"}},"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/huggingface/pytorch-image-models","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/AGI-Arena/MARS","reach":null}],"summary":{"ran_violates":1,"ran_honours":1,"ran_fixture":1,"unverified":2},"by_repo_kind":{"official":{"samples":4,"ran":2,"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":1,"samples":[{"code_sha256_prefix":"aa5486a3650902d8","entry":"exists","repo":null,"repo_kind":null,"path":null,"file_url":null,"link_basis":"identical_code_first_harvested_elsewhere","language":"python","status":"ran_violates","verification_level":1,"contract_check":"VIOLATES","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":null,"inline_ok":false,"mcp_get_code":{"code_sha256":"aa5486a3650902d8"}},{"code_sha256_prefix":"d9145ea1a06f6d52","entry":"get_lr","repo":"AGI-Arena/MARS","repo_kind":"official","path":"MARS/train_mars.py","file_url":"https://github.com/AGI-Arena/MARS/blob/HEAD/MARS/train_mars.py","link_basis":"first_harvest_node","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":"d9145ea1a06f6d52"}},{"code_sha256_prefix":"2cd3b108272e2c4e","entry":"update_fn","repo":"AGI-Arena/MARS","repo_kind":"official","path":"MARS/optimizers/mars.py","file_url":"https://github.com/AGI-Arena/MARS/blob/HEAD/MARS/optimizers/mars.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"2cd3b108272e2c4e"}},{"code_sha256_prefix":"a220f03f6dec9929","entry":"MARS","repo":"AGI-Arena/MARS","repo_kind":"official","path":"MARS/optimizers/mars.py","file_url":"https://github.com/AGI-Arena/MARS/blob/HEAD/MARS/optimizers/mars.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":"a220f03f6dec9929"}},{"code_sha256_prefix":"f155be3896896475","entry":"NewtonSchulz","repo":"AGI-Arena/MARS","repo_kind":"official","path":"MARS/optimizers/mars.py","file_url":"https://github.com/AGI-Arena/MARS/blob/HEAD/MARS/optimizers/mars.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":"f155be3896896475"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}