Hi everyone,

I've trained a Flux Lora model of Hong Kong ICAC Investigators and created a short video using KlingAI.

The training dataset consists of 48 pictures from a TV series, processed over 4,000 steps.

The trigger word is "ICAC", but I recommend to prompt like this:

A photo of ICAC investigator. The image shows a man in a formal suit standing in an office setting. He is wearing a grey suit with a white shirt and a black tie. He has a red lanyard around his neck with a name tag attached to it

A photo of ICAC investigators. The image shows a group of seven young people standing inside an elevator. They are all dressed in formal business attire, with suits and red lanyard around their necks with name tags attached to them.

If you want better results, you can generate images with more loras, as I found that paradiffflux.safetensors is a great one.