Requirements

Python Version

Python >= 3.9

Dependencies

numpy
Pillow
tensorflow>=2.16

Compatibility with recent TensorFlow/Keras releases

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}

Since v1.3.1, this package patches the model loader so exported models load correctly on up-to-date TensorFlow/Keras installs, with no need to pin an old TensorFlow version. See issue #2 for background.