Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
Apriel-H1: Towards Efficient Enterprise Reasoning Models
(
arxiv.org
)
1 point
by
guiriduro
24 days ago
|
hide
|
past
|
favorite
|
1 comment
guiriduro
24 days ago
[–]
Apriel-H1-15b-Thinker-SFT uses incremental distillation from Apriel-Nemotron-15B-Thinker, selectively replacing less critical attention layers with linear Mamba blocks to reduce computational complexity while preserving reasoning quality.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
