image_outpainting有时候会报错，同样的图片

[hjj-lmx]

Wraceback(most recent call last):File "/UD-AI-sD/stable diffusion/server_gpu.py", line 278, in image outpaintingoutputs = brushNet pipeline(params)File "/uD-AI-sD/stable diffusion/power paint/brushNet pipeline.py", line 25, in calloutputs=self.brushNet pipeline(File "/usr/local/lib/python3.10/dist-packages/torch/utils/ contextlib.py", line 1l5, in decorate contextreturn func(*args,**kwargs)File "/uD-AI-SD/stable diffusion/power paint/pipeline/pipeline PowerPaint Brushnet cA.py", line 1453, inlatents = self.scheduler,step(noise pred, t, latents, **extra step kwargs, retum dict-False) [0]File "/usr/local/lib/python3.10/dist-packages/diffusers/schedulers/scheduling unipc multistep .py". line 86sample=self.multistep uni c bh update(File "/usr/local/lib/python3.10/dist-packages/diffusers/schedulers/scheduling unipc multistep.py". line 776Dlt=modelt-mRuntimeError: The size of tensor a (79) must match the size of tensor b (127) at non-singleton dimension

[zengyh1900]

hi @hjj-lmx
你可以在local run 这个看看是什么原因，从报错来看，好像是 tensor size 不对，按道理说同样的图片不应该会出现这个问题