cubiq / ComfyUI_FaceAnalysis

Extension for ComfyUI to evaluate the similarity between two faces
Apache License 2.0
227 stars 25 forks source link

Inhomogeneous shape error in Face Analysis node - Requesting assistance with resolving array shape inconsistency #15

Closed vootox closed 1 month ago

vootox commented 1 month ago

When processing, I get the following error message:

!!! Exception during processing!!! setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (4,) + inhomogeneous part.

I can provide further details if required, I would greatly appreciate any assistance in resolving this issue.

cubiq commented 1 month ago

I need to see the workflow and every reference images

vootox commented 1 month ago

The issue produces a variation (see https://github.com/cubiq/ComfyUI_FaceAnalysis/issues/8#issuecomment-2132481583) of the error with different workflows so I included both workflows in the zip.

Face Analysis Issue.zip

silent37 commented 1 month ago

Hello, have you resolved this issue? I have also encountered the same problem and have upgraded the nodes and versions used to the latest version

cubiq commented 1 month ago

unfortunately I cannot replicate, the provided workflows work as intended for me.

I need the full error backtrace and info about your environment (the initial comfyui message should suffice)

cubiq commented 1 month ago

can you try to update to the latest version I pushed few minutes ago?

vootox commented 1 month ago

unfortunately I cannot replicate, the provided workflows work as intended for me.

I need the full error backtrace and info about your environment (the initial comfyui message should suffice)

I didn't think it was going to be easy to replicate. The only real differences were the input images as snapshots of your video, the workflows are yours.

vootox commented 1 month ago

can you try to update to the latest version I pushed few minutes ago?

Whatever you done to that new code of yours has fixed my problem, it's working! YAY! I actually spent the last 20 hours or so studying tensor shapes with Ollama LLM chat, learning as much as I could to try to understand that part of your code. I learnt I have much to learn!

PS: Even so, I will try various settings and see how thngs go and let you know if unforseen issues arise.