Insights Demo Code - Python Wrapper for Google Teachable Machine

A python based backend for hosting ML models from Google's Teachable Machine and Running demos on webcam

 

This codebase does the following steps:

• Using exported Keras models as REST APIs
• Running exported models at runtime using webcam

Inspiration

This project is greatly inspired from AutoML implementations

Technology

This project uses a number of open source projects to work properly:

• Python - awesome language we love

Environment Setup

Installation

This project requires the standard Python 3.6+ to run (Use 3.6 only, not 3.7 or 3.8)

$ git clone [email protected]:pourabk-hexaware/insights-teachable-machine-wrapper.git

OR

$ https://github.com/pourabk-hexaware/insights-teachable-machine-wrapper.git

$ cd insights-teachable-machine-wrapper
$ pip install -r requirements.txt

Paste download Keras keras_model.h5 and label.txt files undermodel folder

To run webcam demo

$ python webcam.py

To Run Web server

$ python app.py

REST endpoints when running on local

http://localhost:5000/detect for for form-data uploads with key 'image' or,

base64 input { "image_string": "/9....." } as body using header "application/json"

License

Public