What about multiple cloths from one picture
Love the lora
but what about multiple cloths?
I will perform a LoRA for this case, but you can customize it by reducing the weight, using a training version with fewer steps, and increasing the prompt. I think it will be effective.
will you show me a sample by using below picture. I tried with several prompts but it did not work.
Sure, let's wait and see
A quick test. If it works optimally for this case, I think another LoRa chip isn't needed.
can we get a workflow to test more images.
It is entirely possible to build workflows for different objectives and make the LoRA adapt to those specific cases. Or, you can build a branched workflow for each individual scenario.
because mine is not working as expected
Here is more results. or does the lora only works with base
Based on your results, if the customization is not 100% effective, you can add a Ksampler in the second pass and perform the task of removing the body from the clothing.
But I think it can be customized to be effective in a single pass.
But ultimately, LoRA is only trained with data files of single objects. Therefore, I think I will soon train a LoRA for multiple subjects and customize their poses.
Based on your results, if the customization is not 100% effective, you can add a Ksampler in the second pass and perform the task of removing the body from the clothing.
But I think it can be customized to be effective in a single pass.
I tried second pass but it won't work
But ultimately, LoRA is only trained with data files of single objects. Therefore, I think I will soon train a LoRA for multiple subjects and customize their poses.
Well, that’s useful information! Now I can mention it in my YouTube video. A LoRA will be much more effective, since relying on prompts alone doesn't always yield the best results.
But ultimately, LoRA is only trained with data files of single objects. Therefore, I think I will soon train a LoRA for multiple subjects and customize their poses.
Well, that’s useful information! Now I can mention it in my YouTube video. A LoRA will be much more effective, since relying on prompts alone doesn't always yield the best results.
Can i see your Youtube's Channel?