Instructions to use LiconStudio/LTX-2.3-Multiple-Subject-Reference with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use LiconStudio/LTX-2.3-Multiple-Subject-Reference with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("LiconStudio/LTX-2.3-Multiple-Subject-Reference", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
IC-LoRA-Ingredients the new competitor!!
its the new competitor!
https://huggingface.co/Lightricks/LTX-2.3-22b-IC-LoRA-Ingredients
So,how its feeling, I didnt test yet
I think, first of all, there’s no limit on the number of references, because everything is defined within a single sheet.
That makes it easier to control, more compliant with the instructions, and ultimately increases the chances of getting a better output.
I think, first of all, there’s no limit on the number of references, because everything is defined within a single sheet.
That makes it easier to control, more compliant with the instructions, and ultimately increases the chances of getting a better output.
Well, sounds better than MSR
as ingredients pack everything into one sheet, it loses details. MSR wins in many ways
I personally like MSR much better. better character consistency between scenes. also trying to the character sheet is a huge headache so MSR is much easier there as well.
Only thing I did like about ingredients is it had better motion and character facial emotion and less blur on fast and far away shots.
MSR is far easier to use, the prompting requirements to use Ingredients are insane and inconsistent.