Hello @iDron, Indeed, that made a huge difference. After setting the batch size to high numbers (250, 500, 1000), I was able to see the real power of the GPU acceleration. I also found a comment from an engineer of Apple where suggest the same thing. Please refer to this post for the answer, where it suggest that on small batches the training on CPU may be faster.

Hey @TurboVince, I use a virtual env using poetry, and it works fine. This may vary depending on your Python version. Apparently using Python 3.11 to install tensorflow-macos 2.12 does not work yet (Not sure why). Apart from that, everything works great. GPU is available after installing tensorflow-metal, and there are some limitation when using optimizers such as tf.keras.optimizers.Adam. Let's give it a try and we can see how that works.