Skip to content

An unofficial inference process demonstration Gradio application for DiffusionDet.

License

Notifications You must be signed in to change notification settings

XIRZC/diffusiondet.demo

Repository files navigation

Demonstration App for DiffusionDet with Denoising Process Visualization

This is just a simple demo application built by Gradio.

DiffusionDet Gradio Interface

Environment Preparation

You can just refer to DiffusionDet for basic environment preparation, where we just add gradio package additionally.

Specifically,

conda create -n diffusiondet python=3.7
conda activate diffusiondet
pip install 'git+https://ghproxy.com/https://github.com/facebookresearch/detectron2.git'
pip install gradio
# In Linux, prepare model weights by
mkdir models && wget https://ghproxy.com/https://github.com/ShoufaChen/DiffusionDet/releases/download/v0.1/diffdet_coco_swinbase.pth -P ./models/
# In Windows, prepare model weights manually

Otherwise, you can just use our docker image hosted on AliYun, and the correponding Dockerfile is located at here.

docker pull registry.cn-hangzhou.aliyuncs.com/mrxir/diffusiondet.demo:model-executable-v2.0

Usage

By local environment:

  • In Linux, just run ./gradio_demo.sh under this repo root directory.
  • In Windows, just run python gradio_demo.py under created diffusiondet conda virtual environment

By docker container:

Just run docker run -it --rm -p 7860:7860 registry.cn-hangzhou.aliyuncs.com/mrxir/diffusiondet.demo:model-executable-v2.0 in any OS terminal with docker installed.

After deploy the gradio web application, you can just enter your localhost in browser.

About

An unofficial inference process demonstration Gradio application for DiffusionDet.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages