Skip to content

Latest commit

 

History

History
61 lines (36 loc) · 1.92 KB

README.md

File metadata and controls

61 lines (36 loc) · 1.92 KB

SIGE Interactive SDEdit Demo

Overview

SIGE achieves 2x less conversion time compared to original DDPM on M1 MacBook Pro GPU as we selectively perform computation on the edited regions.

Prerequisites

  • Python3
  • CPU, M1 GPU, or NVIDIA GPU + CUDA CuDNN
  • PyTorch >= 1.7. For M1 GPU support, please install PyTorch>=2.0.

[Notice] Our code is tested on M1 MacBook Pro with PyTorch 2.0. However, it should be runnable on CUDA and CPU machines.

Getting Started

Setup

  • Install PyTorch. To reproduce our CUDA and CPU results, please use PyTorch 1.7. To enable MPS backend, please install PyTorch>=2.0.

  • Install PyQt5. On M1 MacBook Pro, it can be installed with Conda:

    conda install pyqt

  • Install SIGE following ../README.md. Remeber to set the environment variables if you are using M1 GPU.

  • Install other dependencies:

    conda install tqdm -c conda-forge
pip install pyyaml easydict gdown

Running

  • Original DDPM

    python start.py --config_path configs/church_dpmsolver256-original.yml

  • SIGE DDPM

    python start.py --config_path configs/church_dpmsolver256-sige.yml

By default, these commands will test results on GPU if GPU is available. You can also explicitly specify the device with --device. If the model downloading is too slow for you, you can switch the download source from our website to Google Drive with --download_tool gdown.

Acknowledgement

This frontend is developed based on Piecasso. The backend is developed based on SDEdit, ddim and dpm-solver.