welter
@fun
interested in contracting ai dev work for good $$$ goal: - create comfyui workflow - looking for consistent character generation from 1 single image requirements: - input "image_url" and "prompt" and output a new image - base model MUST be flux 1.1 pro ultra raw (or as close as possible) if you or someone you know is proficient in comfyui and is confident you/they can accomplish this for me, i will pay a good few grand for this
10 replies
6 recasts
54 reactions
Conway
@conway
might be able to help you but curious why you don’t want to fine tune a model and exactly what type of character since it depends on context
1 reply
0 recast
0 reaction
welter
@fun
for my usecase i would just need to input a photo of a person having to finetune a model for every time i want to create another photo of a person would be frustrating
1 reply
0 recast
0 reaction
Conway
@conway
you’re probably going to either need to get over that hurdle or be okay using masking the problem with using image input is if you want it to look different than the image, without masking, that’s going to get applied everywhere
1 reply
0 recast
0 reaction
welter
@fun
if masking can get me 95% the way there and be passable, im good with using that if i gave you example input images, would you want to take a stab at this?
2 replies
0 recast
0 reaction
Conway
@conway
btw fwiw, besides collecting samples, finetuning isn’t hard so if you do have access to multiple high quality shots of the person, definitely recommend it since you’d be able to do a lot very easily once you have the lora
0 reply
0 recast
0 reaction
Conway
@conway
I’d need example inputs and some description of output goals but I can give you a quick sense of how easy it’d be for us to compare to other people’s workflows
1 reply
0 recast
0 reaction
