Revision History
Gradient Accumulation · 10 revisions
Sizes are character counts of the article source. The signed number is the change from the previous revision.
Recent edit summaries
Detailed summaries recorded by editors. Generic maintenance summaries are omitted here; every recorded revision remains below. Dates describe the edit, not necessarily the event it covers.
- Correction: Corrected the 7B-model memory figure (about 112 GB of model state under ZeRO's 16-bytes-per-parameter accounting, not 56 GB), a Llama 2 batch-size calculation that was off by a factor of four, and a claim that FSDP needs a higher learning rate than DeepSpeed, which Hugging Face traced to a precision-handling difference (plus 13 minor fixes)
Version 10 · Nov 2, 2026, 04:46 PM
- Improve internal links (2026-10-13)
Version 9 · Oct 13, 2026, 10:08 AM
- Added 6 contextual internal links
Version 8 · Jul 23, 2026, 03:05 PM