Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
You Only Cache Once: Decoder-Decoder Architectures for Language Models
(
arxiv.org
)
3 points
by
reqo
26 days ago
|
hide
|
past
|
favorite
|
1 comment
simonttt
22 days ago
[–]
Basically, it is a hybrid model like Jamba.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
