Could you clarify the differences between the following models
#6
by
makisekurisu-jp
- opened
relighting-kontext-dev-lora-v3.safetensors,
重新打光.safetensors,
and the converted relight-kontext-lora-single-caption_comfy.safetensors?
The filenames don’t make it clear, and I’d like to understand how they differ in terms of training data, usage, or compatibility.
The root directory of the folder contains Qwen-Edit's LoRa dataset, which was trained using the images and prompts from the folder I uploaded.
The folders in the Kontext directory contain datasets collected from the internet; as mentioned in my description, I am not aware of the datasets they used.