search

comfyanonymous / ComfyUI

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
https://www.comfy.org/
GNU General Public License v3.0
56.61k stars 6k forks source link

Input types 'tensor<1x257x1xf16>' and 'tensor<1xf32>' are not broadcast compatible #3084

Open Emvolatile1 opened 8 months ago

Emvolatile1 commented 8 months ago

Hello, I am not very elaborate on the topic and I would really appreciate if someone helps me out on the following issue(I am running on M2 Pro Mac):

got prompt model_type EPS Using split attention in VAE Using split attention in VAE clip missing: ['clip_l.logit_scale', 'clip_l.transformer.text_projection.weight'] Requested to load CLIPVisionModelProjection Loading 1 new model loc("varianceEps"("(mpsFileLoc): /AppleInternal/Library/BuildRoots/ce725a5f-c761-11ee-a4ec-b6ef2fd8d87b/Library/Caches/com.apple.xbs/Sources/MetalPerformanceShadersGraph/mpsgraph/MetalPerformanceShadersGraph/Core/Files/MPSGraphUtilities.mm":233:0)): error: input types 'tensor<1x257x1xf16>' and 'tensor<1xf32>' are not broadcast compatible LLVM ERROR: Failed to infer result type(s).

yaoGreat commented 5 months ago

you can add "--force-fp32" in cmdline, like this: python main.py --force-fp32 and it works for me. https://huggingface.co/CompVis/stable-diffusion-v1-4/discussions/13 it seems there is some issuse of float64