Flux Text-to-Image Workflow Guide: Which Version is Best for You? (ComfyUI workflows included)

[@RobinHahnRN]

Thanks, Sharvin - excellent, detailed video as always.

[@CodeCraftersCorner]

Thank you!

[@TheAutisticRebel]

Your the best buddy. Thank you for the ridiculous detail!!! 🎉🎉🎉 You Rock!

[@CodeCraftersCorner]

Glad to help

[@ilyanemihin6029]

Thank you very much for the information on where to download the T5-XXL and CLIP files! A couple of days ago, it took me about an hour of searching on Google to find out what they were and where to get them)

[@CodeCraftersCorner]

Thank you!

[@001mcv]

Another well crafted excellent educational video and a great set of resources sure to be useful to many! Thank you!

[@CodeCraftersCorner]

Glad you enjoyed it!

[@SebAnt]

Thanks Sharvin.

[@CodeCraftersCorner]

Thank you Sebastian!

[@suraiverse]

Great explanation dude as usual . Thanks for sharing

[@CodeCraftersCorner]

Thank you!

[@alecubudulecu]

Great video. And great breakdown in the workflows. Love this

[@CodeCraftersCorner]

Glad you enjoyed it!

[@klb1113]

Thanks so much, Sharvin! I really appreciate your detailed explanation of things and your openly sharing the workflows. One thing: how can one add Loras to these workflows, for those of us who still find ComfyUI confusing? Can you share this information or do a quick demo of what to do/where to insert it? If that's too much to ask, I certainly understand and still thank you for your informative videos.

[@xcom9648]

You can insert load lora nodes in between load model/dualcliploader and clip text encode/modelsamplingflux.

[@klb1113]

@@xcom9648 Thanks! I will give that a try!

[@Ollegruss_Music]

Thanks. Spasibo!

[@CodeCraftersCorner]

Thanks.

[@WhySoBroke]

Thanks for this great information!

[@CodeCraftersCorner]

Glad it was helpful!

[@murphylanga]

Thank you, very well explained and summarized, I will link it again under my latest video.

[@CodeCraftersCorner]

Glad it was helpful!

[@7arso-w3e]

Thanks, Sharvin , pls can you talk about the new InstantIR and how to install it?

[@CodeCraftersCorner]

Okay, I will try.

[@erlinghagendesign]

hi Sharvin, excellent. The images on Github? Not yet uploaded?

[@CodeCraftersCorner]

Oops! I’d missed uploading them. They’re up on GitHub now. Thanks for letting me know.

[@bahramkhadiv2234]

Hi, Thanks indeed for your detailed tutorial. I am an architect and looking for a best options for sketch to render workflows by Flux and segmentation(or similar) with Flux to start switching all basic 3D geometries into detailed realistic images. It would be appreciated if you suggest or provide a tutorial for that too.

[@CodeCraftersCorner]

Hello @bahramkhadiv2234, thank you for your comment. I searched for architectural sketches and found mostly technical designs floor plans. I don't think we can work with these in flux. However, hand-drawn or digital outlines of buildings might work better. Could you help me find some reference images I can use to test some workflow? Even keywords to search on google will work.

[@MilesBellas]

👍👍👍

[@CodeCraftersCorner]

Thank you!

[@giuseppedaizzole7025]

Hi, nice video. I have a question. for dev original you put load diffusers and for dev fp8 load checkpoint. why dev original is not on checkpoint or viceversa? i'm using mostly Forge, lately im trying Confy and i share my checkpoint folder from forge but there is not diffuser folder to share with. so....is there a way to make work dev original placed in forge? ...there are 2 questions...lol....thanks

[@CodeCraftersCorner]

The original dev checkpoint contains only the model component, so we need to add the clips and the vae separately. The fp8 version contains all three components, so we can use the load checkpoint node for it. You can create the folder manually and put the models in there. It should not affect you stable diffusion forge and comfyui will pick up the new folder automatically.

[@giuseppedaizzole7025]

@@CodeCraftersCorner i'll check it, thanks