Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
moffkalast
on Dec 20, 2023
|
parent
|
context
|
favorite
| on:
LLM in a Flash: Efficient LLM Inference with Limit...
That's what the Google TPU is in a nutshell as I understand it, loading weights into memory cells between fpus.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
