teachable-machine-lite 1.2.0.2

A lightweight Python package optimized for integrating exported models from Google's Teachable Machine Platform into robotics and embedded systems environments. This streamlined version of Teachable Machine Package is specifically designed for resource-constrained devices, making it easier to deploy and use your trained models in embedded applications.

Teachable Machine Lite

By: Meqdad Darwish

Description

A lightweight Python package optimized for integrating exported models from Google's Teachable Machine Platform into robotics and embedded systems environments. This streamlined version of Teachable Machine Package is specifically designed for resource-constrained devices, making it easier to deploy and use your trained models in embedded applications. With a focus on efficiency and minimal dependencies, this tool maintains the core functionality while being more suitable for robotics and IoT projects.

Source Code is published on GitHub

Read more about the project (requirements, installation, examples and more) in the Documentation Website

Supported Classifiers

Image Classification: Use exported and quantized TensorFlow Lite model from Teachable Machine Platform (a model file with tflite extension).

Requirements

For detailed information about package requirements and dependencies, please visit our documentation

Python >= 3.9

numpy < 2.0 (v1.26.4 recommended)

How to install Teachable Machine Lite Package

pip install teachable-machine-lite

Dependencies

numpy
tflite-runtime
Pillow

Example

An example for teachable machine lite package with OpenCV:

from teachable_machine_lite import TeachableMachineLite
import cv2 as cv

cap = cv.VideoCapture(0)

model_path = "model.tflite"
labels_path = "labels.txt"
image_file_name = "screenshot.jpg"

tm_model = TeachableMachineLite(model_path=model_path, labels_file_path=labels_path)

while True:
    ret, img = cap.read()
    cv.imwrite(image_file_name, img)

    results, resultImage = tm_model.classify_and_show(image_file_name, convert_to_bgr=True)
    print("results:", results)

    cv.imshow("Camera", resultImage)
    k = cv.waitKey(1)
    if k == 27:  # Press ESC to close the camera view
        break

cap.release()
cv.destroyAllWindows()

Values of results are assigned based on the content of labels.txt file.

For more; take a look on these examples

Links:

• Documentation

• PyPI

• Source Code

• Teachable Machine Platform