Hacker News
new
top
best
ask
show
job
Ask HN: What's the best LLM model that on a 24 GB VRAM GPU?
3 points
by
max93
4 hours ago
2 comments
jr_isidore
2 hours ago
Good question. I was told in 2024 to get an RTX 3090 (24 GB VRAM), so I did, and nothing on HuggingFace was usable.
roscas
3 hours ago
I would love to see Laguna.xs-2 on that, because it's very good on cpu only.