360View Redmond Z-Image Turbo Panorama LoRA
LoRA for Z-Image Turbo that generates 360-degree panoramic scenes: cityscapes, aurora borealis, volcanic landscapes, and immersive wide environments.
360View Redmond Z-Image Turbo Panorama LoRA
A LoRA adapter for Tongyi-MAI/Z-Image-Turbo that pushes the model toward wide-format 360-degree panoramic compositions. Instead of standard frames, it generates immersive scenes suited for environment previews, concept backgrounds, and panoramic placeholders.
- Trigger with
360or360 VIEW.in the prompt. - Best results in ComfyUI with the Z-Image Turbo base model.
- Works across urban, arctic, and volcanic environments.
- Apache 2.0 licensed.
What it does
This LoRA teaches Z-Image Turbo a panoramic bias: wider field of view, curved spatial reading, and scene-scale subjects. The sample outputs include a night cityscape with dense illuminated towers, an aurora borealis view over a frozen landscape, and a volcanic terrain with flowing lava. Those examples suggest the model preserves detail across the full 2:1 panorama aspect rather than collapsing edges.
How to use
Load it as a LoRA on top of Tongyi-MAI/Z-Image-Turbo in ComfyUI. Include 360 or 360 VIEW. in the prompt to activate the style. The model card does not list exact weight or CFG values, so start with standard Z-Image Turbo settings and adjust LoRA strength if the panorama framing feels too weak or too distorted.
Recommended settings
- Base model: Tongyi-MAI/Z-Image-Turbo
- Trigger words:
360,360 VIEW. - Aspect ratio: wide or 2:1 panoramic
- Interface: ComfyUI
- License: Apache 2.0
Common errors and fixes
Error | Cause | Fix | Source
|---|---|---|---
Weak panorama effect | LoRA weight too low or missing trigger | Raise LoRA strength slightly; add 360 VIEW. | HF README Edge distortion | Extreme wide aspect with default sampler | Reduce aspect ratio slightly or use more denoising steps | Observed in samples Missing safetensors load | Wrong file format | Use the .safetensors weights from the model repo | HF README
FAQ
Does it work in AUTOMATIC1111? The card recommends ComfyUI, but the LoRA format is standard; it should load in other UIs that support diffusers LoRA weights.
Do I need all three trigger words? Use 360 or 360 VIEW.; both are listed as triggers.
Can it do non-panorama images? It is tuned for panoramic output, so standard-framed results may be less consistent.
Sources
- HuggingFace model card for artificialguybr/360VIEW-REDMOND-ZIMAGETURBO
- Redmond.AI GPU sponsorship noted in the model description