Hello,This is my first Lora created with FluxGym on the default settings. It was created from the surviving miniature souvenir views of Ballingrath Gardens in Mobile, Alabama, and the city of Natchez, Mississippi. It captures the photopainting style of the training material and transfers it to an existing image or helps in generating prompted content. I would strongly suggest checking the training material for deeper insight into the used Florence 2-based custom captioning and what the base material looks like to have some awareness of outputs that are to be expected. For concepts in some way connected with the training material, I would suggest setting the strength of the Lora at 0.7 minimum, while for more loosely associated ideas, I would go up for 1.0 to a maximum of 1.5. I'm also providing a workflow under the link <a target="_blank" rel="ugc" href="https://civarchive.com/articles/10349">https://civarchive.com/articles/10349</a> that I've used to achieve the different gen qualities presented in the model showcase. The workflow saves each gen after each corresponding KSampler, giving 3 versions in total. Because this is my first Lora I have made some sloppy cropping on a few of the training views, and the Lora rarely likes to generate, in some cases, a slim, white strip on one of the sides of the gen. Sorry for that. If someone has more of those miniature postcards in color and would wish to add this set to his own training material in the intention of creating a photopainting lora of his own or would just like to add this small collection to his own training project, please feel free to do that. If one intends to blatantly publish or redistribute this Lora without crediting the user Sebik420 account, I would wish for the individual that it would shrivel, dry up, and fall off from between his legs. :P

Files with SHA256: 86fe49702075e198d38921da318b73836f767c7b9f463a4fa109e8fdb368916d