OS: Windows 10 (Surface Pro4) No GPU
--Download and install Miniconda from the following site
- https://conda.io/miniconda.html
--Use Miniconda3-latest-Windows-x86_64.exe
--As of 06/10/2018, conda4.5.4 and python3.6.5 are installed by default.
--Create a virtual environment for deep learning with conda and install the necessary modules
--Launch Anaconda Prompt from the Start Menu
--Create virtual environment: conda create -n env_name python = 3.5
--Create with the above command assuming python3.5, which has many library support
--Virtual environment activation: ʻactivate env_name --Installation of required modules: Many modules can be installed from each conda pip, but priority is given to conda --From conda:
conda install numpy scipy sympy pandas scikit-learn matplotlib seaborn pillow ipython cython pytest jupyter m2w64-toolchain Werkzeug pyyaml tensorflow keras --Install from pip:
pip install ipdb`
--Operation check
--Select Create New Project and create a project in any location and with any name.
--If you are asked for Project Interpreter at the time of creation, specify (install_path) \ Miniconda3 \ envs \ env_name \ python.exe
--Changes after creation: From File> Settings> Project: project_name> Project Interpreter
--Update of what you installed with conda: conda update --all
--Update what you installed with pip: pip install module_name -u