Do you have magic prompts?
Hi, thanks for sharing the LoRA.
I was wondering if you used any kind of “magic prompt” or specific prompt template for style transfer when training or testing it.
If so, would you mind sharing an example?
Hi, thanks for sharing the LoRA.
I was wondering if you used any kind of “magic prompt” or specific prompt template for style transfer when training or testing it.
If so, would you mind sharing an example?
It uses fixed prompt words, as detailed in the description, and no other descriptive words were used for training.
Hi, thanks for answering.
I sometimes run into a case where the model seems to confuse the input image and the condition image. For example, instead of transferring the condition image’s style to the input image, it sometimes appears to mix up their roles, or the condition image ends up dominating the result.
Have you seen this issue before? If so, I’d really appreciate any tips on how to handle it, such as prompt wording, image order, or parameter settings.
Hi, thanks for answering.
I sometimes run into a case where the model seems to confuse the input image and the condition image. For example, instead of transferring the condition image’s style to the input image, it sometimes appears to mix up their roles, or the condition image ends up dominating the result.Have you seen this issue before? If so, I’d really appreciate any tips on how to handle it, such as prompt wording, image order, or parameter settings.
To avoid style remnants or errors, the input image must be a line drawing, while the reference image is a colored style drawing.