Instructions to use Lightricks/LTX-2.3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use Lightricks/LTX-2.3 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline from diffusers.utils import load_image, export_to_video # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Lightricks/LTX-2.3", dtype=torch.bfloat16, device_map="cuda") pipe.to("cuda") prompt = "A man with short gray hair plays a red electric guitar." image = load_image( "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/guitar-man.png" ) output = pipe(image=image, prompt=prompt).frames[0] export_to_video(output, "output.mp4") - Notebooks
- Google Colab
- Kaggle
Sharing: LTX2.3 Subtitles and Removal !
Thanks for releasing LTX-2-3, but the generated videos often include subtitles, some of which have spelling errors or garbled Chinese characters.
Therefore, is there an option in LTX-2-3 to completely disable subtitle output?
I'm hoping for a way to completely prevent subtitle generation.
Thanks!
subtitle in negative prompt.
But you figured that out in other thread ;-)
So for anyone else that see this
hate that too, neg.prompt still not work (sometimes it works, weird...), translate that into Chinese might help maybe.
hate that too, neg.prompt still not work (sometimes it works, weird...), translate that into Chinese might help maybe.
For negative prompts to work you need CFG higher than 1, or if you are on ComfyUI use NAG node
KJNodes have LTX Nag node for ComfyUI
Can see example of NAG with CFG 1 (aka using LTX distilled model workflows) here:
https://huggingface.co/Kijai/LTX2.3_comfy/discussions/10#69aacedfbf85f3a773a8e3ff
Based on my testings this due to euler_anc samplers. Res_2s both stages should fix this!