When using cnvrg to run experiments or to start a notebook, cnvrg fires up a machine with many different libraries and setting that come pre-installed for the use of data science.

There are some cases when a user would like to use a different version of a library or a library that isn't installed on the machine.

To support that, cnvrg offers an easy solution as it searches for a requirements.txt file and automatically installs its content prior the run of the experiment/notebook.

Before running a new experiment or starting a notebook:

  • Create a file name 'requirements.txt' and locate it in the project head tree (in the main projects directory)
  • specify which python libraries you would like to install
  • If you want to use the web interface for running experiments or starting notebooks:make sure to sync the file before starting an experiment or a notebook with cnvrg sync
  • If you use cnvrg CLI tool for running experiments or starting notebooks, the file will be synced automatically

When running the experiment or starting a notebook, cnvrg will first locate the requirements.txt file, will identify which environment to run it against (Python 2 or Python 3) and only then will start the experiment or notebook session.

An example of an experiment with requirements.txt file

docutils==0.11
Jinja2==2.7.2
MarkupSafe==0.19
Pygments==1.6
Sphinx==1.2.2
numpy
tensorflow

To generate this kind of format to your own project, type 

$ pip freeze > requirements.txt

An example of project tree with requirements.txt:

$ ls
app/
src/
requirements.txt

 

 

 

 

 

 

Did this answer your question?