Hey guys, just a quick release of my FLUX workflow
-It includes text to image and image to image. Also has a wildcards option setup.
-Also has image prompting where you load an image and let an LLM caption it and generate an image based off the caption (moondream or florence2 setup already)
-Upscaling + Color correction + Image Comparison Node.
Description
This one fixes the latent interpolation on the image to image. You will have to download (via comfy manager) the new node I created to do this: https://github.com/NeuralSamurAI/ComfyUI-Dimensional-Latent-Perlin
I also upgraded the options to edit parameters of the conditioning and guidance. Enjoy.
FAQ
Comments (14)
I'm getting an error about a missing node NoisyLatentPerlinD I have a noisylatentperlin node but it is missing the input for latent image.
Addressed in the model info
"About this version
This one fixes the latent interpolation on the image to image. You will have to download (via comfy manager) the new node I created to do this: https://github.com/NeuralSamurAI/ComfyUI-Dimensional-Latent-Perlin"
Same here. Missing NoisyLatentPerlinD
Addressed in the model info
"About this version
This one fixes the latent interpolation on the image to image. You will have to download (via comfy manager) the new node I created to do this: https://github.com/NeuralSamurAI/ComfyUI-Dimensional-Latent-Perlin"
Maybe just write a little guide of which toggles to be on and off for each objective, especially for image2image
This one is better for IMG2IMG
https://civitai.com/models/625887/simple-and-effective-flux1-img2img-upscale-comfyui-workflow
Turn on img2img. You can leave off the image prompting nodes, upscale node, both of those are optional.
The fast muter does all the work for you, and the context switches keep the data correct. So just flip it on or off and thats all you need to do!
Maybe you should test your workflow before releasing. If you turn off parts of the workflow like upscale it dosen't save the image correctly, givers the old upscaled image. Also Moondream is missing an output so it wont work either???
Hook the moondream output to your text input instead of the positive prompt. I will be releasing a new version shortly that addresses those issues you mentioned.
Maybe you should rely less on workflows others make and learn it yourself. Its much faster than commenting and waiting for someone else to update. I do this for free and there is absolutely no guarantee anything will work. If thats not your cup of tea, plenty of other workflows out there ;)
@harmonicdiffusion I have the moondream workflow working, but im not the one releasing this workflow
@admajic glad you have it working! enjoy! I dont release wf's for others. these are what I use myself. You can take it or leave it, im not trying to sell anything, so i really dont care =)
@admajic Chill dude. He shared this with you for free. Honestly, an unreal amount of entitlement.
@harmonicdiffusion thank you for sharing!
@ryanontheinside Huh? 3 months later why comment under my comment? I dunno people hate Comfyui because its complicated. Makes it even harder when you give them a broken workflow. But if you think receiving broken workflows is the way then enjoy this broken workflow and good luck ;P
When disabling i2i, the wf complains about height/width not set. any fix ?