Gimbal360

Single-image 360° panorama generation from an arbitrary perspective photograph.

Code: https://github.com/Orange-3DV-Team/Gimbal360

Model Weights

File	Size	Description
autolevel.pth	107 MB	Geometry estimation model combining a pretrained SegNeXt-B backbone with a trained flow decoder and rigid filter. Estimates camera FOV, pitch, and roll from a single perspective image.
pytorch_lora_weights.safetensors	171 MB	Fine-tuned panorama-generation LoRA weights for FLUX.1-Fill-dev.

Usage

The inference script downloads both weights automatically from this repository:

git clone https://github.com/Orange-3DV-Team/Gimbal360.git
cd Gimbal360
pip install -r requirements.txt

python inference.py \
  --image example/example.png \
  --prompt "a frozen, abandoned train car in a post-apocalyptic wasteland." \
  --resolution 960

Weights can also be fetched manually:

hf download Orange-3DV-Team/Gimbal360

Pipeline

1. AutoLevel (autolevel.pth) estimates the camera parameters (FOV / pitch / roll) of the input photo.
2. The photo is projected onto an equirectangular canvas as conditioning, with an inpaint mask for the unknown regions.
3. FLUX.1-Fill-dev + LoRA (pytorch_lora_weights.safetensors) inpaints the full 360° panorama.

License

The LoRA weights are a derivative of FLUX.1-Fill-dev and are released under the FLUX.1-dev Non-Commercial License.