Yet another inflation LoRA, this time fine-tuned on Qwen Image 2.1. Still didn't get around to that dataset expansion, soz. Been lazy and spending more time playing video games. White Knuckle is really good.
Anyways, I'm not a huge fan of Qwen Image 2.1 for T2I. But for editing, it's the open SOTA (we'll see about that again when FLUX.3 comes out, I guess.) I only trained on basic text-image pairs, but it works well for editing too.
Qwen Image 2.1 is already decent at expansion, but it didn't go big enough for my tastes, hence this.
(In case it needs to be said: Don't use this to violate people's consent. If you do and I find out, I'm taking this down and helping whoever it was take action against you. Don't be that guy.)
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support