2. Getting Started
2.1. Installation from PyPI
MLPro is listed in the Python Package Index (PyPI) and can be installed using the package installer for Python (pip) in two variants:
Without any dependencies
The following command installs the latest version of MLPro.
pip install mlproAdditional packages may need to be installed manually (depending on the functionalities you intend to use).
Full installation with all dependencies
There is also an option to automatically install MLPro and all depending packages in validated versions (see Subsection Dependencies below). This option will ensure that all the functionalities of MLPro, including wrappers and examples, work appropriately out of the box.
pip install mlpro[full]
2.2. Installation from Anaconda
MLPro is also available on Anaconda and can be installed with the following command:
conda install -c mlpro mlpro
2.3. Dependencies
The table below shows all packages that MLPro has dependencies on. Additionally, the versions with which MLPro is compatible are listed. Since we cannot influence incompatible changes on dependent packages, we unfortunately cannot rule out the possibility of problems occurring with different versions. We review and update the list with each new release.
Which packages are actually required depends on the functionalities of MLPro that are used.
Package |
Version |
---|---|
dill |
0.3.6 |
numpy |
1.24.2 |
torch |
2.0.0 |
matplotlib |
3.7.1 |
scipy |
1.10.1 |
multiprocess |
0.70.14 |
pandas |
2.1.3 |
2.4. First Steps
The easiest way to become familiar with the concepts and functions of MLPro is to browse through the numerous example programs. We can also recommend taking a closer look at the key features of MLPro and following the links.