Loop Forge

Same park at strength 1.0

Limit, from LTX 2.5 Creative Lab LoRAs, tested locally

Frame from the Day source result
Day source
Frame from the Day-to-Night, 1.0 result
Day-to-Night, 1.0
Watch it in the video at 7:17
LoRA
Day-to-Night at 1.0
Output
1152×640, 121 frames at 24fps
Sampling
8 steps, euler_ancestral
Seed
42
Generated in
9 min
Peak VRAM
11.0 GiB
Ran on
RTX 3060, 12GB (local)

What was wired in

Limit

At 1152×640 the default strength only lights the lamps. The scene stays partly in daylight, with some parts of it day and some night.

A realistic nighttime version of the same park: the fountain, bench, lamp post and trees in the same positions and the jogger following the same path, the lamp post lit with a warm glow, cool moonlight on the gravel and the fountain water catching the light, deep natural shadows under the trees. Only the lighting changes from day to night; identical composition, framing, camera movement and motion.

Notes

The LoRA was trained at 768×448 with 97 frames. At the larger size used here it needed a strength of 1.3 to 1.5.

The Day-to-Night LoRA on Hugging Face

Run it yourself

Download everything for this test (.zip, 1.9 MB). It holds the workflow, every input it loads and a README. Copy the inputs into ComfyUI/input, put the models below in place and drag the workflow onto ComfyUI.

The Lightricks repositories (the LTX 2.5 base model and the LoRAs) are gated. Log in to Hugging Face and accept the terms on each repository page before downloading.

Custom nodes: ComfyUI-LTXVideo at commit 3bf3ca6. Everything else is part of ComfyUI 0.37.0 (commit 0f74f7fb, frontend 1.53.6).