
This Lora is trained on 9 Midjourney images that were influenced by the 1972 Rothschild Surrealist Ball. For the past year or so I've struggled to recreate these images with SDXL and even Midjourney itself. I was a bit blown away by how well this Lora turned out with so little training images. Flux really is joy to work with.
Just describe the people you want at the party and what you want them to wear. You don't really need to do much beyond that prompting wise. The Lora will do the rest. If you don't describe the characters doing something with their hands, Flux tends to generate the same hand positions due to the lack of training images. The same with the faces, you may need describe facial features to get variation.
Sampler: Eular
Steps: 30
Guidance: 3.5
描述:
训练词语:
名称: lora.TA_trained.safetensors
大小 (KB): 598391
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success