"Small Batch Size Training for Language Models: When Vanilla SGD Works, and ..."

Martin Marek et al. (2025)

> Home

Details and statistics

DOI: 10.48550/ARXIV.2507.07101

access: open

type: Informal or Other Publication

metadata version: 2025-08-12

- view
  - electronic edition via DOI (open access)
  authority control:
- export record
  dblp key:
  - journals/corr/abs-2507-07101
- ask others
- share record
  persistent URL:
  - https://dblp.org/rec/journals/corr/abs-2507-07101
Martin Marek, Sanae Lotfi, Aditya Somasundaram, Andrew Gordon Wilson, Micah Goldblum:
Small Batch Size Training for Language Models: When Vanilla SGD Works, and Why Gradient Accumulation Is Wasteful. CoRR abs/2507.07101 (2025)

manage site settings

To protect your privacy, all features that rely on external API calls from your browser are turned off by default. You need to opt-in for them to become active. All settings here will be stored as cookies with your web browser. For more information see our F.A.Q.