Hardworking laborers, there's work to do. The Nunchaku wheel has been updated to v1.2.0.
Hardworking laborers, there's work to do. The Nunchaku wheel has been updated to v1.2.0.
yes plz. that would be great
Hi, thank you very much for your feedback. Yes, I saw the new version was added this morning, and I think I'll add it :)
Given that I have to make some changes on my computer, you'll have to be a little patient, especially since the compilations take a long time.
Hi, thank you very much for your feedback. Yes, I saw the new version was added this morning, and I think I'll add it :)
Given that I have to make some changes on my computer, you'll have to be a little patient, especially since the compilations take a long time.
Thank you
the wheels have been added.
the wheels have been added.
Thank you,
the wheels have been added.
Thank you. god bless you ^^
2.9.0 + cu130 when?
2.9.0 + cu130 when?
This version is already available from the official nunchaku repository. I only added the missing versions.
See nunchaku repo
Hi ,
first of all I would like to thank you so much for taking this opportunity to make the missing versions of Nunchaku ,it's really appreciable ! <3
Though, I tried it but didn't work for me, it ends up with the following error:
Node `NunchakuFluxDiTLoader` import failed:
Traceback (most recent call last):
File "ComfyUI\custom_nodes\ComfyUI-nunchaku\__init__.py", line 82, in <module>
from .nodes.models.flux import NunchakuFluxDiTLoader
File "ComfyUI\custom_nodes\ComfyUI-nunchaku\nodes\models\flux.py", line 16, in <module>
from nunchaku import NunchakuFluxTransformer2dModel
File ".venv\Lib\site-packages\nunchaku\__init__.py", line 1, in <module>
from .models import (
...<6 lines>...
)
File ".venv\Lib\site-packages\nunchaku\models\__init__.py", line 1, in <module>
from .text_encoders.t5_encoder import NunchakuT5EncoderModel
File ".venv\Lib\site-packages\nunchaku\models\text_encoders\t5_encoder.py", line 18, in <module>
from .linear import W4Linear
File ".venv\Lib\site-packages\nunchaku\models\text_encoders\linear.py", line 10, in <module>
from ..._C.ops import gemm_awq, gemv_awq
ImportError: DLL load failed while importing _C: The specified module is not found.
knowing that:
- I tried to install the following version: nunchaku-1.2.0+torch2.7-cp313-cp313-win_amd64.whl
- Venv python version is: 3.13.9
- Cuda version is: V11.8.89
can you please help me ? thanks a lot !
Hi ,
first of all I would like to thank you so much for taking this opportunity to make the missing versions of Nunchaku ,it's really appreciable ! <3Though, I tried it but didn't work for me, it ends up with the following error:
Node `NunchakuFluxDiTLoader` import failed: Traceback (most recent call last): File "ComfyUI\custom_nodes\ComfyUI-nunchaku\__init__.py", line 82, in <module> from .nodes.models.flux import NunchakuFluxDiTLoader File "ComfyUI\custom_nodes\ComfyUI-nunchaku\nodes\models\flux.py", line 16, in <module> from nunchaku import NunchakuFluxTransformer2dModel File ".venv\Lib\site-packages\nunchaku\__init__.py", line 1, in <module> from .models import ( ...<6 lines>... ) File ".venv\Lib\site-packages\nunchaku\models\__init__.py", line 1, in <module> from .text_encoders.t5_encoder import NunchakuT5EncoderModel File ".venv\Lib\site-packages\nunchaku\models\text_encoders\t5_encoder.py", line 18, in <module> from .linear import W4Linear File ".venv\Lib\site-packages\nunchaku\models\text_encoders\linear.py", line 10, in <module> from ..._C.ops import gemm_awq, gemv_awq ImportError: DLL load failed while importing _C: The specified module is not found.knowing that:
- I tried to install the following version: nunchaku-1.2.0+torch2.7-cp313-cp313-win_amd64.whl
- Venv python version is: 3.13.9
- Cuda version is: V11.8.89
can you please help me ? thanks a lot !
Hello, thanks for your feedback!
Regarding the issue, the version of your CUDA drivers is far too old (CUDA 11.8).
I compiled the wheels using CUDA 12.9, so you’ll need to install the latest NVIDIA drivers, which you can download directly from the NVIDIA website.
Once that’s done, everything should work correctly .
You’ll just need to make sure you’re using a compatible version of PyTorch. I recommend torch 2.8+cu128.
Hello, thanks for your feedback!
Regarding the issue, the version of your CUDA drivers is far too old (CUDA 11.8).
I compiled the wheels using CUDA 12.9, so you’ll need to install the latest NVIDIA drivers, which you can download directly from the NVIDIA website.
Once that’s done, everything should work correctly .
You’ll just need to make sure you’re using a compatible version of PyTorch. I recommend torch 2.8+cu128.
Hi!
I would like to thank you very very much for your support! actually I followed your recommendation and now it worked for me very well with the updated NVidia Driver & Cuda 12.9 as well as torch 2.8+cu128 .
Big thanks for you !!