FreeReg运行笔记
1. 环境配置
- 下载代码
git clone https://github.com/WHU-USI3DV/FreeReg.git
- 创建conda环境
conda env create -f environment.yaml
- 安装MinkowskiEngine
下载MinkowskiEngine
git clone https://github.com/NVIDIA/MinkowskiEngine.git
安装依赖
sudo apt install libopenblas-dev
安装MinkowskiEngine
python setup.py install --blas_include_dirs=${CONDA_PREFIX}/include --blas=openblas
- 下载MiDaS
在终端中开一个python,或者把代码写为脚本
import torch
torch.hub.help("intel-isl/MiDaS", "DPT_BEiT_L_384", force_reload=True) # Triggers fresh download of MiDaS repo
记录MiDaS下载的路径,然后把该路径填入midas = torch.hub.load()中
2. 运行
python demo.py
可能遇到如下问题
2.1. 缺少openai/clip-vit-large-patch14
安装
sudo apt install git-lfs
初始化
git lfs install
下载(Unpacking运行时间很长)
git clone https://huggingface.co/openai/clip-vit-large-patch14
2.2. 缺少stable-diffusion-v1-5
下载(时间很长)
git clone https://huggingface.co/runwayml/stable-diffusion-v1-5
参考文献
GitHub - WHU-USI3DV/FreeReg: [ICLR 2024] FreeReg: Image-to-Point Cloud Registration Leveraging Pretrained Diffusion Models and Monocular Depth Estimators