I noticed in your documentation that the prompts are designed to include a class_word. Could you please clarify the range of acceptable class_word values? Are they limited to terms like "man", "woman", "boy", and "girl"? Additionally, if I aim to create a more general prompt that is inclusive of both genders, would "person" be an appropriate class_word to use?
Additionally, I'd like to share an observation from my experiments: removing the background from the input image appears to enhance the preservation of the ID. Is there an explanation for that?
Hi there, thanks for your excellent work!
I noticed in your documentation that the prompts are designed to include a class_word. Could you please clarify the range of acceptable class_word values? Are they limited to terms like "man", "woman", "boy", and "girl"? Additionally, if I aim to create a more general prompt that is inclusive of both genders, would "person" be an appropriate class_word to use?
Additionally, I'd like to share an observation from my experiments: removing the background from the input image appears to enhance the preservation of the ID. Is there an explanation for that?
Looking forward to your insights!