282 B
282 B
title, excerpt, link, date, tags
| title | excerpt | link | date | tags |
|---|---|---|---|---|
| DeepSpeed Ulysses: System Optimizations for Enabling Training of Extreme Long Sequence Transformer Models | https://github.com/deepspeedai/DeepSpeed/blob/master/blogs/deepspeed-ulysses/README.md | 2023-08-24 00:00:00 | training ZeRO English |