Floydhub's environments come preinstalled with most of the common deep learning and machine learning packages and dependencies. Examples of pre-installed packages include numpy, scipy, OpenCV, OpenAI Gym, SpaCy, etc. 

In the case when you need additional packages, you can install them before running your job or restarting your Workspace. You can install two types of packages:

Why do I need to reinstall the package for every time I run the workspace?

Unfortunately, the packages you install in a Job  or Workspaces is not preserved when the run finishes. We are currently planning to support preserving of environment changes in the near future.

Advanced workaround

There is a workaround for this
. You can create a FloydHub dataset which contains the changes to the filesystem from installing dependencies. Then you can mount this dataset inside the Workspace and copy the dependencies in the right path (take a look at the script used by this user for this workaround: create_floyd_base.sh). 

Let us know if you will need more help by reaching out to us at [email protected] or via the Intercom chat bubble.

Did this answer your question?