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") - LTX-2
How to use Lightricks/LTX-2.3 with LTX-2:
# Install the LTX-2 pipelines git clone https://github.com/Lightricks/LTX-2.git cd LTX-2 uv sync --frozen
# Download the weights from this repo, plus the Gemma text encoder hf download Lightricks/LTX-2.3 --local-dir models/LTX-2.3 hf download google/gemma-3-12b-it-qat-q4_0-unquantized --local-dir models/gemma-3-12b
# Fast pipeline (distilled model, no distilled LoRA needed) uv run python -m ltx_pipelines.distilled \ --distilled-checkpoint-path models/LTX-2.3/<distilled-checkpoint>.safetensors \ --spatial-upsampler-path models/LTX-2.3/<spatial-upsampler>.safetensors \ --gemma-root models/gemma-3-12b \ --prompt "A beautiful sunset over the ocean" \ --output-path output.mp4 # For image-to-video, add: --image path/to/image.jpg 0 0.8# HQ pipeline (two-stage, higher quality) uv run python -m ltx_pipelines.ti2vid_two_stages_hq \ --checkpoint-path models/LTX-2.3/<checkpoint>.safetensors \ --distilled-lora models/LTX-2.3/<distilled-lora>.safetensors 0.8 \ --spatial-upsampler-path models/LTX-2.3/<spatial-upsampler>.safetensors \ --gemma-root models/gemma-3-12b \ --prompt "A beautiful sunset over the ocean" \ --output-path output.mp4 # For image-to-video, add: --image path/to/image.jpg 0 0.8 - Notebooks
- Google Colab
- Kaggle
Lightsabers ? lora ?
When I use LTX 2.3 to create videos with lightsabers the lightsaber hilts are turned into weird thick objects although I provide a good "first frame" with star wars style lightsaner hilts. why is that ?
can I use a lora to improve lightsabers ? how ?
did you try LTX 2.5 ?
I was not aware that 2.5 was out. I tried it now.
It has definitely better and acceptable lightsaber hilt understanding.
But I see other problems with it.
First; I used the I2V workflow from here:
https://civitai.com/models/2852094/ltx25-basic-workflow-t2v-i2v
I have 1 input image. The image shows 8 people dressed in jedi clothes and holding 1 ligthsaber each.
My prompt to the workflow is:
"8 jedi characters raising their lightsaber slightly towards the camera in sync. there should not be any other people in the video.
Each character is holding his/her own lightsaber and the color of each lightsaber must be same throughout the video. Characters keep their natural expression throughout the video.
They are each holding one lightsaber in just one hand.
There are only eight characters in the video. Only the characters from the first image.
Camera does not zoom in."
The result is a nice video with some issues:
- lightsabers have sparkles around the blade
- characters are laughing
How can I resolve these issues ?