Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
brookman64k
6 months ago
|
parent
|
context
|
favorite
| on:
DeepSeek-R1
Thank you. Which is currently the most capable version running reasonably fast on a 3090 (24GB of VRAM)?
danielhanchen
6 months ago
[–]
The Llama distilled version Q4_K_M should be reasonably fast and good!!
Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: