GSFix3D
Collection
Diffusion model collections for paper "GSFix3D: Diffusion-Guided Repair of Novel Views in Gaussian Splatting" • 4 items • Updated • 2
import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("goldoak1421/gsfixer-full", dtype=torch.bfloat16, device_map="cuda")
prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]GSFixer Full Model Card
This model card is for the gsfixer-full model, which is designed for novel view repair of an image rendered from a 3DGS model with additional information gain from the mesh. The model is fine-tuned from the stable-diffusion-2 model following Marigold protocol.
Other Versions
- gsfixer-base: A version of this model that operates solely on 3DGS input.
Model Details
- Developed by: Jiaxin Wei, Stefan Leutenegger, Simon Schaefer.
- License: CreativeML Open RAIL++-M License.
- Model Description: This model can be used to generate a repaired image of two input rendered images from the mesh and 3DGS, respectively.
- Resolution: Even though any resolution can be processed, the model inherits the base diffusion model's effective resolution of roughly 768 pixels. This means that for optimal predictions, any larger input image should be resized to make the longer side 768 pixels before feeding it into the model.
- Steps and scheduler: This model was designed for usage with the DDIM scheduler and between 1 and 50 denoising steps.
- Cite as:
@article{wei2026gsfix3d, title={GSFix3D: Diffusion-Guided Repair of Novel Views in Gaussian Splatting}, author={Wei, Jiaxin and Leutenegger, Stefan and Schaefer, Simon}, booktitle={2026 International Conference on 3D Vision (3DV)}, year={2026}, organization={IEEE} }
- Downloads last month
- 11
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support