Skip to content
Better HN
Top
Best
Ask
Show
New
Jobs
Search
⌘K
1.8-3.3x faster Embedding finetuning now in Unsloth
(opens in new tab)
(unsloth.ai)
3 points
electroglyph
5mo ago
3 comments
Save
Share
3 comments
3 comments · 3 top-level
top
newest
oldest
storystarling
5mo ago
Do the memory savings carry over to inference or is this strictly optimizing the backward pass? I'm running embedding pipelines via Celery and being able to squeeze this into lower VRAM would help the margins quite a bit.
danielhanchen
5mo ago
Excited to have collabed on this! Thanks electroglyph for the contrib!
electroglyph
OP
5mo ago
see also:
https://www.reddit.com/r/LocalLLaMA/comments/1qk9vmv/1833x_f...
j
/
k
navigate · click thread line to collapse