Hacker Times
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
SlavikCA
70 days ago
|
parent
|
context
|
favorite
| on:
Qwen3.6-27B: Flagship-Level Coding in a 27B Dense ...
There is no standard.
The higher quantization - the better results, but more memory is needed. Q8 is the best.
SV_BubbleTime
70 days ago
[–]
FP32 is best, although I wonder if there isn’t something better I don’t know about. Q8 is for the most part equal to FP16 in practical terms by being smart about what is quantized, but iirc always slower than FP16 and FP8.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
The higher quantization - the better results, but more memory is needed. Q8 is the best.