streamlit-tfx: TensorFlow Extended visualizers for Streamlit apps

streamlit-tfx provides utilities for visualizing TensorFlow Extended artifacts in Streamlit apps.

🌱 Just sprouting!

This project is in the very beginning stages of development. It's not well tested and is only intended to be used as a demo.

Installation

pip install streamlit-tfx

Getting started

import streamlit_tfx as st_tfx

st_tfx.display(item)
st_tfx.display_statistics(statistics)
st_tfx.display_schema(schema)
st_tfx.display_anomalies(anomalies)
st_tfx.display_eval_result_plot(eval_result)
st_tfx.display_eval_result_slicing_attributions(eval_result)
st_tfx.display_eval_result_slicing_metrics(eval_result)
st_tfx.display_eval_results_time_series(eval_results)

Most artifacts in tests/artifacts/ were generated by running the TFX Keras Component tutorial. The anomalies artifact with anomalies was generated by running the TensorFlow Model Analysis tutorial.

🚀 Inspired by spacy-streamlit and streamlit-player.