Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
modeless
on Feb 24, 2023
|
parent
|
context
|
favorite
| on:
LLaMA: A foundational, 65B-parameter large languag...
How possible is it to run these models on a gaming GPU?
sp332
on Feb 25, 2023
|
next
[–]
If you're patient,
https://github.com/FMInference/FlexGen
lets you trade off GPU RAM for system RAM or even disk space.
coolspot
on Feb 24, 2023
|
prev
|
next
[–]
Yes, if your gaming rig is NVidia DGX-1 workstation.
https://youtu.be/5TRr2oWeSw0
_kuvn
on Feb 25, 2023
|
prev
[–]
a very rough approximation is 2GB of vram for every billion fp16 model parameters, so the lower end models may be just about achievable on high-end cards
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
