Running tensorflow data validation (facets and dive) on Streamlit

Streamlit is really cool. I know there are some threads on HTML and security to execute in Streamlit. I was wondering if I can embed Tensorflow based components such as TFDV and TFMA within a running Streamlit app.

Has anyone gotten this to work? I can do this in a Jupyter Notebook with iPython but unsure if we can do in Streamlit. Thanks!

Hi @ahsanshah, thanks for posting. We do not explicitly support it. It seems TFDV and TFMA are based on facets. I added a feature request to investigate how to add it:

We also have a plugin architecture coming up that may enable more specialized extensions to be added to Streamlit:


1 Like