VACE

VACE copied to clipboard

vace_wan_inference.py gets killed during pipeline creation with 1.3B model on G5.xlarge EC2 instance

4

I'm encountering an issue where the vace/vace_wan_inference.py script gets killed during the inference process while using the vace-1.3B model. The termination happens immediately after the "Creating WanT2V pipeline" log message....

Shivanshu8211

Is there a VACE 'recipe' for using a video for the reference image?

1

I ask because: * Right now, I can add a static image of a room in a house, and ask Wan/VACE to "put a dog in the middle of the...

AILova

Inpainting with pose control

2

Hello, I have been trying to generate videos using inpainting + pose control, but I have found that the generated results are not normal. The following img is one generated...

Jxzh2020

Repainting outpainting of video size and duration

1

For repainting outpainting, I use Wan2.1-VACE-1.3B model to inference. The run script is below torchrun \ --nproc_per_node=2 \ --master_port=29500 \ vace/vace_pipeline.py \ --base wan \ --task outpainting \ --direction 'up,down,left,right'...

wenxin27

@hanzhn Thanks for great work! We recently tackle the core challenges of ​​Subject-to-Video Generation (S2V)​​ by systematically building the first complete infrastructure—featuring an evaluation benchmark and a million-scale dataset! ✨Welcom...

SHYuanBest

VACE 14B runs out of memory on an H100 with 80GB of VRAM.

2

VACE 14B causes an out-of-memory error on an H100 with 80GB of VRAM. How much VRAM is required to run it on a single GPU?

guruguruzom

Does Wan2.1 model support distributed inference?

3

Does the Wan2.1 model support multi-GPU distributed inference? If yes, how to set parameters or adjust the code? Wan2.1模型支持多Gpu推理吗？如何设置参数或调整代码呢？

wenxin27

Image guided video to video example?

1

Hi, as above, is there example code or a cli command to run image guided video to video \video restyling, directly, without comfyui? Thanks!

antithing

Different skeletons/structures

1

Hi, I'm trying to use VACE for non-human structures as input. Is that possible? Say a dog/bear or a more complex animal like a dragon with wings. I can create...

lucas-vitrus

运行 python .\vace\gradios\vace_wan_demo.py 报错 RuntimeError: Failed to import transformers.models.clip.modeling_clip because of the following error (look up to see its traceback): No module named 'triton'

1

RuntimeError: Failed to import transformers.models.clip.modeling_clip because of the following error (look up to see its traceback): No module named 'triton' 有人遇到过吗

anstonjie