InstantX
InstantX
Here is [something](https://twitter.com/radamar/status/1749592253138653595) I found online.
Working on it.
It works now on [hf](https://huggingface.co/spaces/InstantX/InstantID).
Thanks for all your interests. As @antonioo-c has suggested, we do have plan to release code, so, no worries. However, the code needs to go through some internal approvals before...
Ah, it is a template. It's ready now! @h3clikejava
Do you mean you want use a reference image to guide the style, instead of prompt only? For now, using IP-Adapter with CLIP is a choice.
We will update soon
Besides of Pytorch 2.0. We have already added an example for [LCM-LoRA](https://github.com/InstantID/InstantID?tab=readme-ov-file#speed-up-with-lcm-lora), which allows for fewer steps and can improve speed quite a bit.
We will add a img2img pipeline soon. But you can also do it on your own and PR for us, as our pipeline is fully compatible with existing diffusers pipelines.
We don't use WebUI's plugins.