Spaces:
Running
on
A10G
Apply for community grant: Academic project (gpu)
This demo space complements our paper titled "PatchFusion: An End-to-End Tile-Based Framework for High-Resolution Monocular Metric Depth Estimation" by Zhenyu Li, Shariq Farooq Bhat, Peter Wonka.
PatchFusion extends ZoeDepth to high-resolution input, providing high-resolution depth estimation maps with intricate details. This opens up vast number of possibilities. For example, combining with ControlNet, it achieves better depth-guided text-to-image generation. Some of the examples produced via the Gradio UI can be seen at https://zhyever.github.io/patchfusion/
This demo would benefit greatly from GPU hardware for better interactivity and user experience.
We provided a temporary example demo here: https://55510c1c829b28b9e3.gradio.live/, and observed many people were using it. Limited by resource, it would be much better if we could get your help!
Hi @zhyever , we have assigned a gpu to this space. Note that GPU Grants are provided temporarily and might be removed after some time if the usage is very low.
To learn more about GPUs in Spaces, please check out https://huggingface.co/docs/hub/spaces-gpus
Thank you for the help. I'm checking it now. It might be caused by the default high-resolution.
Ok, done. It's inevitable to reduce processing resolution for controlnet. I will clarify the overall pipeline in introduction then. Thanks for the help!