Feature Idea
Add support for the official https://huggingface.co/Lightricks/LTX-2.3-22b-IC-LoRA-Ingredients's in-context references system.
Required input format
The lora requires a reference sheet that's a single image (or combined set of images fit into the video resolution as a single image) identical across every frame.
The prompt requires 2 sections, Reference sheet: and Generated video:, but this could be written by the user.
Suggested implementation
When reference images are used with LTX 2/2.3 in the prompt box, they should be joined together and used as every frame of the video. If videos ever get support to be added to a prompt those could also work for in context video-to-video, like the other ltx creative lab loras.
Other
No response
Feature Idea
Add support for the official https://huggingface.co/Lightricks/LTX-2.3-22b-IC-LoRA-Ingredients's in-context references system.
Required input format
The lora requires a reference sheet that's a single image (or combined set of images fit into the video resolution as a single image) identical across every frame.
The prompt requires 2 sections,
Reference sheet:andGenerated video:, but this could be written by the user.Suggested implementation
When reference images are used with LTX 2/2.3 in the prompt box, they should be joined together and used as every frame of the video. If videos ever get support to be added to a prompt those could also work for in context video-to-video, like the other ltx creative lab loras.
Other
No response