Overview

TF Watcher is a Python package and web app that allows users to monitor their Machine Learning training or testing process on mobile devices. It is specifically designed for use with Google Colab, Azure ML, and Kaggle. The project includes two subprojects: TF Watcher Python Package and TF Watcher Web App. The Python package is built using TensorFlow and Pyrebase, allowing users to easily monitor the metrics they want and write them to a Firebase realtime database. The web app is built using React, Chakra-UI, Recharts, and Firebase, and it displays the logs using charts.

Features

• Simple to use Python package
• Designed for monitoring Machine Learning training or testing process on mobile devices
• Built for Google Colab, Azure ML, and Kaggle
• TF Watcher Python Package uses TensorFlow and Pyrebase
• TF Watcher Web App uses React, Chakra-UI, Recharts, and Firebase

Installation

To install TF Watcher Python Package, follow these steps:

1. Navigate to the /tfwatcher directory.
2. Execute the following command: pip install tfwatcher

To install TF Watcher Web App, follow these steps:

1. Navigate to the /webapp directory.
2. Execute the following command: npm install tfwatcher

Summary

TF Watcher is a useful tool for monitoring Machine Learning processes on mobile devices. The Python package and web app offer easy integration with popular platforms such as Google Colab, Azure ML, and Kaggle. The Python package uses TensorFlow and Pyrebase to write metrics to a Firebase realtime database, while the web app displays logs using charts. TF Watcher is a valuable resource for ML developers who want to keep track of their training and testing progress.