Nuclei Segmentation Link:
https://jgamper.github.io/PanNukeDataset/
see the notes: win10_python3.5(keras)_python3.7(torch).docx if you want to install two pythons on win10, and one python using keras, another using torch
(1) win10+pycharm+install geopandas, where I encountered the errors to install by using pip, the soluions are: https://stackoverflow.com/questions/34427788/how-to-successfully-install-pyproj-and-geopandas
Under windows
As @Marcus Williams states:
Download Fiona and GDAL Windows Binaries corresponding your python environment (32, 64bit and 2.7 3.3 etc. (from http://www.lfd.uci.edu/~gohlke/pythonlibs/) and go in cmd with cd to directory where downloaded files are and install the .whl files using the command
pip install filename.whl Finally:
pip install geopandas
In my case: I downloaded: GDAL‑3.0.4‑cp37‑cp37m‑win_amd64.whl, Fiona-1.8.13-cp37-cp37m-win_amd64.whl
first install GDAL, then Fiona, finally geopandas
(2) in my lab win10 computer with python3.7, I installed pysal 2.0.0 successfully
(1) The shell to access linux machine: (i) SSH Secure Shell Client (ii) MobaXterm
(2) CPU, GPU, SSD -> speeding up deep learning
https://jgamper.github.io/PanNukeDataset/
(1) download the github files from the website: https://github.com/facebookresearch/SparseConvNet
(2) run bash develop.sh, it may report the error: OSError: CUDA_HOME environment variable is not set. Please set it to your CUDA install root. solve the problem by the following the solution: https://blog.csdn.net/OOFFrankDura/article/details/113632416 一次性:在命令行 export CUDA_HOME=/usr/local/cuda-X.X
(3) it may report the error that gcc version should be greater than 5, then following the solution in the web: https://www.cnblogs.com/jixiaohua/p/11732225.html 1、安装centos-release-scl
sudo yum install centos-release-scl 2、安装devtoolset，注意，如果想安装7.版本的，就改成devtoolset-7-gcc，以此类推
sudo yum install devtoolset-8-gcc* 3、激活对应的devtoolset，所以你可以一次安装多个版本的devtoolset，需要的时候用下面这条命令切换到对应的版本
scl enable devtoolset-8 bash