Getting Started¶
Prerequisites¶
It is encouraged to use environment wrapper and package manager, conda is chosen as the reference solution. Please follow the installation section in their official guide.
Some of the codes require CUDA dependency, please download the binary release from the NVIDIA website.
Quick start¶
Create an empty workspace named demo, if you have your preferred environment, feel free to skip the following two lines. It is highly recommended to use numpy and scipy from conda instead of pip, since their version has mkl support directly.
conda create -n demo python=3 numpy scipy
conda activate demo
Install uToolbox using pip
pip install utoolbox
or try out pre-release by
pip install --pre utoolbox
Congratz! You can now try out stuff in the guides section.