SD3 Image generation workflow with upscaling and metadata. Now with updated conditioning nodes, face detailer, image comparison and preview chooser.
This uses the full size SD3_medium model with clips and t5. If you are using the smaller versions you will also need the separate clip encoders and you can load them using a triple clip loader.
Description
FAQ
Comments (12)
Where do you get the Vae? The very last step you have a SD3_Medium_Vae ?
I can't load the model, I get this error:
Dtype not understood: F8_E4M3
I love this workflow 😻
Can you post a list of the extensions and their versions that you are using? I can't get working versions of all the extensions necessary for this to work
Awesome. trying it now :)
Error occurred when executing CLIPTextEncode: 'NoneType' object has no attribute 'tokenize'
Thank you MilitantHitchhiker! Great work!
Error occurred when executing SDPromptSaver: expected str, bytes or os.PathLike object, not NoneType
This is pretty similar to a workflow I was using - close enough I can use this as a starting point with minimal changes :)
How do you get rid of the cartoon aspect? I used my usual SD prompts and still cartoons :( I even prompted photograph of
I've never had this problem with SD3 workflows before, so I figure I got one of the many settings wrong :(
Can someone please explain why we are sending the negative prompts in this workflow at short intervals with all these conditioning and not directly?
Very handy but I can't move any of your nodes. What's up with that?
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.