Testing Lora Captioning for Styles
Dataset of 6 Images, not very diverse flash photos, trained via CivitAI-RAPID Flux training
Versions:
no Caption
single Caption (flashphoto)
joy Caption
joy Caption with Trigger (flashphoto, [joyCaption natural language] - not explicity trained on trigger since i used civitAI Trainer)
Showcase with and without "Trigger"
Description
FAQ
Comments (1)
I've used your excellent 1.5 and SDXL models before. On this one though, in my testing, the 'flash style' is good but it seems to override my prompt with similar looking females - so I can't really use it unfortunately.