search

yardenfren1996 / B-LoRA

Implicit Style-Content Separation using B-LoRA
MIT License
302 stars 22 forks source link

Question about multi images training #19

Closed yatoubusha closed 4 months ago

yatoubusha commented 4 months ago

Thank you for your answer about issue #18, If I want to use multiple images to train content Lora, do I need to make any changes to the training code besides preparing the dataset?

yardenfren1996 commented 4 months ago

No @yatoubusha , you don't need any additional changes except for the prompt. To get better results (since we don't need to keep the structure of the image), you should use the prompt "A [v] {class_name}" as suggested by the authors of Dreambooth in their paper.