By: Meqdad Darwish
A Python package designed to simplify the integration of exported models from Google's Teachable Machine platform into various environments. This tool was specifically crafted to work seamlessly with Teachable Machine, making it easier to implement and use your trained models.
Source Code is published on GitHub
Read more about the project (requirements, installation, examples and more) in the Documentation Website
Image Classification: use exported keras model from Teachable Machine platform.
Teachable Machine's exported .h5 models embed a legacy DepthwiseConv2D layer config that current Keras (Keras 3, bundled by default since TensorFlow 2.16) rejects with an error such as:
TypeError: Unrecognized keyword arguments passed to DepthwiseConv2D: {'groups': 1}
Some exports also save the model as a Sequential wrapping nested Sequential/Functional submodels, a shape Keras 3's legacy H5 loader mis-rebuilds, which previously surfaced as a misleading FileNotFoundError: Model file not found.
Since v1.3.1, this package patches the model loader to handle both cases, so exported models load correctly on up-to-date TensorFlow/Keras installs, with no need to pin an old TensorFlow version. It also fixes prediction-annotation crashes on Windows / recent Pillow versions (show_prediction_on_image). See issue #2 and the changelog for background.
Python >= 3.9
pip install teachable-machineAn example for teachable machine package with OpenCV:
from teachable_machine import TeachableMachine
import cv2 as cv
cap = cv.VideoCapture(0)
model = TeachableMachine(model_path="keras_model.h5",
labels_file_path="labels.txt")
image_path = "screenshot.jpg"
while True:
_, img = cap.read()
cv.imwrite(image_path, img)
result, resultImage = model.classify_and_show(image_path)
print("class_index", result["class_index"])
print("class_name:::", result["class_name"])
print("class_confidence:", result["class_confidence"])
print("predictions:", result["predictions"])
cv.imshow("Video Stream", resultImage)
k = cv.waitKey(1)
if k == 27: # Press ESC to close the camera view
break
cap.release()
cv.destroyAllWindows()Values of result are assigned based on the content of labels.txt file.
For more; take a look on these examples