Loop Forge
Loop Forge / Projects / LTX 2.5 Creative Lab LoRAs, tested locally

Town square, camera arc

Limit, from LTX 2.5 Creative Lab LoRAs, tested locally

Frame from the Clay layout (Blender) result
Clay layout (Blender)
Frame from the Look still (Qwen-Image) result
Look still (Qwen-Image)
Frame from the Layout to Render result
Layout to Render
Watch it in the video at 16:16
LoRA
Layout to Render at 1.0
Output
1920×1088, 121 frames at 24fps
Sampling
8 + 3 steps, euler_ancestral
Seed
42
Generated in
36 min
Peak VRAM
11.8 GiB
Ran on
RTX 3060, 12GB (local)

What was wired in

  • Clay layout
    Guide video
    Clay layoutRendered in Blender from primitives with clay shadingDownload MP49.0 MB
  • Look still
    Look stillQwen-Image 2.1 edit of the layout's first frame, with a character photoDownload PNG4.2 MB
Limit

I built the layout in Blender from primitives and made the look still from its first frame with Qwen-Image 2.1. The camera arc, the fountain, the tree and the lamp land where the layout put them. The box mannequin has no front, so she turns her back mid-shot. The nearest bench smears, and the empty sky in the layout became more houses.

The prompt

Download workflowDownload everything for this test (.zip, 13.2 MB) 
A young woman with long curly red hair, in a cream knitted sweater and blue jeans, stands beside a weathered sandstone fountain in a small European town square at golden hour as the camera slowly arcs around her and moves closer. Cream and ochre townhouses with shuttered windows, a café awning, a leafy plane tree, black iron street lamps, wooden benches and worn cobblestones, warm low sunlight with long soft shadows. Match the first frame. Photoreal.

Notes

I have seen more side effects like these on busier layout animations, so I expect it to work better with simpler camera moves. The prompt matches the look still and leaves out the words clay and 3D.

The Layout to Render LoRA on Hugging Face

Run it yourself

Download everything for this test (.zip, 13.2 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.

FilePut it inSize
ltx-2.5-22b-distilled-transformer-comfy-int8-convrot.safetensors gatedmodels/diffusion_models21.5 GB
gemma4-12b-with-proj-ltx-2.5-comfy-int8-convrot.safetensors gatedmodels/text_encoders15.4 GB
ltx-2.5-audio-vae-bf16.safetensors gatedmodels/vae364.9 MB
ltx-2.5-video-vae-bf16.safetensors gatedmodels/vae1.5 GB
ltx-2.5-latent-spatial-upscaler-x2-bf16-1.0.safetensors gatedmodels/latent_upscale_models995.8 MB
ltx-2.5-22b-ic-lora-layout-to-render-1.0.safetensors gatedmodels/loras1.3 GB

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).