teneva-opti 0.6.0

Collection of various optimization methods, including tensor based, for multivariate functions and multidimensional data arrays

teneva_opti

Description

Collection of various optimization methods (search for the global minimum and/or maximum) for multivariate functions and multidimensional data arrays (tensors). This library is based on a software product teneva. See also related benchmarks library teneva_bm.

Installation

1. The package can be installed via pip (it requires the Python programming language of the version 3.8 or 3.9):

   pip install teneva_opti==0.6.0
   

   The package can be also downloaded from the repository teneva_opti and be installed by python setup.py install command from the root folder of the project.

2. We test optimizers with benchmarks from teneva_bm library. For installation of additional dependencies (gym, mujoco, etc.), please, do the following (for existing conda environment teneva_opti; if you are using a different environment name, then please make the appropriate substitution in the script; note that you don't need to use environment in colab):

   wget https://raw.githubusercontent.com/AIRI-Institute/teneva_bm/main/install_all.py && python install_all.py --env teneva_opti && rm install_all.py
   

   In the case of problems with scikit-learn, uninstall it as pip uninstall scikit-learn and then install it from the anaconda: conda install -c anaconda scikit-learn. If you have problems downloading the script via wget, you can download it manually from the root folder of the repository teneva_bm.

Documentation and examples (in progress...)

Please, run the demo script from the root of the teneva_opti repository:

clear && python demo/base.py

See also other demo scripts in the folder demo of the teneva_opti repository.

Authors

• Andrei Chertkov
• Gleb Ryzhakov
• Ivan Oseledets

---

✭__🚂 The stars that you give to teneva_opti, motivate us to develop faster and add new interesting features to the code 😃