Skip to content

Latest commit

 

History

History
32 lines (26 loc) · 2.38 KB

README.md

File metadata and controls

32 lines (26 loc) · 2.38 KB
Raw

deepchem-for-dummies

My experiences as a chemist for installing and using deepchem

As a medicinal chemist interested in deep learning and trying to understand new tools to increase productivity I was very interested when I saw the blogpost by Derek Lowe on his in the pipeline blog about Deepchem. Since I am not a programmer I was a bit overwhelmed when I tried to use the tool and was originally unable to get any of the tutorials to work. After a little bit of experience I have been at least able to get the tutorials to work.

Here I plan on documenting first the installation of Deepchem and the required dependencies then as I get more familiar with the software I plan on documenting a how to use deepchem for dummies.

Installation

  • Install a working version of Linux
    • I used Linux Mint installed on Hyper-V on a Windows 10 computer at work. It would work even better on a native Linux box. Since Mint is a fork of Ubuntu/Debain I will use those commands
  • Install Miniconda - I used miniconda so I could control which packages/dependencies I install
    • download the installer, then navigate to the ~/Downloads folder in the terminal
    • bash Miniconda3-latest-Linux-x86_64.sh
  • Install git and download the deepchem repository - I built from the repository for 2 reasons. 1) it worked 2) so I would have the notebooks
  • Install Deepchem in a conda environment called deepchem
    • $ cd deepchem
    • $ bash scripts/install_deepchem_conda.sh deepchem
    • $ conda activate deepchem
    • $ python setup.py install
  • Install the required dependencies and Jupyter notebook - I used conda then updated/upgrades with pip. Make sure you are within your deepchem environment
    • $ conda install numpy, pandas, joblib, scikit-learn, tensorflow, matplotlib
    • $ conda install -c conda-forge rdkit
    • $ pip install --upgrade tensorflow
    • $ conda install jupyter
  • You should now have all of the required libraries to use deepchem - open the jupyter notebook and navigate to the deepchem notebooks
    • $ jupyter notebook