SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2411.10438 · 2024

MARS: Unleashing the Power of Variance Reduction for Training Large Models

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 5 functions out of this paper's own repositories and ran 3 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
AGI-Arena/MARS canonical 2 of 4
copy not recorded — 1 of 1
FunctionStatusWhere it lives
exists Ran this paper's copy was not recorded; identical code first harvested from ThomasMrY/VCT
pointer only · get_code("aa5486a3650902d8")
get_lr Ran AGI-Arena/MARS/MARS/train_mars.py
code served (permissive licence) · get_code("d9145ea1a06f6d52")
update_fn Ran AGI-Arena/MARS/MARS/optimizers/mars.py
code served (permissive licence) · get_code("2cd3b108272e2c4e")
MARS Not yet run AGI-Arena/MARS/MARS/optimizers/mars.py
code served (permissive licence) · get_code("a220f03f6dec9929")
NewtonSchulz Not yet run AGI-Arena/MARS/MARS/optimizers/mars.py
code served (permissive licence) · get_code("f155be3896896475")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

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.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2411.10438")
get_code_for_paper("2411.10438")
have("2411.10438")

Connect an agent — have() is free.