Skip to content

Issues: mit-han-lab/nunchaku

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

27 Open 27 Closed
27 Open 27 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

Support loading local svdquant model in ComfyUI
#56 opened Dec 13, 2024 by wardensc2
SVDQuant Quality Issue: text vs object position, finger number
#55 opened Dec 12, 2024 by marvin-0042
2
HUNYUAN video model Quants? when? 😁
#53 opened Dec 11, 2024 by 4lt3r3go
LyCORIS support
#52 opened Dec 11, 2024 by JakobLS
Details about weight preprocessing scripts, smoothing factor determination, weight decoupling, and quantization code
#51 opened Dec 11, 2024 by FTanghu
Unable to pin memory: operation not supported
#49 opened Dec 10, 2024 by booksword
3
Your method corrupts text-generation alignment
#48 opened Dec 9, 2024 by Oguzhanercan
9
Please write a more detailed install guide on how to install in ComfyUI
#47 opened Dec 8, 2024 by wardensc2
10
Install Error
#46 opened Dec 6, 2024 by jhss
5
LORA switch feature
#44 opened Dec 5, 2024 by QZH-eng
4
How can I modify the model to support multiples of 8 to accommodate the sizes I need?
#43 opened Dec 4, 2024 by kelisiya
3
do you think you could support HunyuanVideo the new video model from tehcent?
#42 opened Dec 3, 2024 by Manni1000
4
Even if it does not download the model, initialization takes 48gb ram.
#40 opened Dec 1, 2024 by vgabbo
6
Pip package
#38 opened Nov 28, 2024 by adhikjoshi
1
Precision used in multi head attention layer
#37 opened Nov 26, 2024 by kimdaeun00
When set DEBUG = True in setup.py, I get the following error:
#36 opened Nov 25, 2024 by Lenan22
4
New Flux Fill Support
#34 opened Nov 22, 2024 by chuck-ma
2
Implementation dtails: GEMM_W4A4::quantize behavior different from the paper
#33 opened Nov 22, 2024 by xmfbit
Is that possible to support any torch model?
#30 opened Nov 21, 2024 by microbenh
1
Python-level Linear Modularization.
#24 opened Nov 14, 2024 by Artiprocher
1
Can nunchaku support more Video-generator model？
#23 opened Nov 14, 2024 by philipwan
6
Running 'example. py' in a Docker container can cause the system to freeze or be forced to shut down
#22 opened Nov 13, 2024 by ziyaxuanyi
11
Question about why getting blank image when using flux ControlNet.
#20 opened Nov 13, 2024 by chuck-ma
7
How to quantize a custom Flux model? For example, lightweight models like Flux Lite that have removed some double blocks.
#13 opened Nov 11, 2024 by kelisiya
2
RuntimeError: Error compiling objects for extension
#12 opened Nov 11, 2024 by 973398769
7
ProTip! Add no:assignee to see everything that’s not assigned.