Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
xg15
33 days ago
|
parent
|
context
|
favorite
| on:
AirLLM 70B inference with single 4GB GPU
If token speed doesn't count, you can have 200GB "RAM" in swap space.
But I agree, what the OP does is a lot more efficient than this.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
But I agree, what the OP does is a lot more efficient than this.