The move to DJL v0.29.0 means that TensorFlow now works on Apple Silicon.
I've tested this with the StarDist extension, and the speed improvement is pretty major (about 4-5x on m M1 Max).
We should therefore remove the warning about TensorFlow on Apple Silicon - but, at the same time, check what happens with TensorFlow and PyTorch on older Intel Macs, in case we need to introduce new warnings there.
The move to DJL v0.29.0 means that TensorFlow now works on Apple Silicon.
I've tested this with the StarDist extension, and the speed improvement is pretty major (about 4-5x on m M1 Max).
We should therefore remove the warning about TensorFlow on Apple Silicon - but, at the same time, check what happens with TensorFlow and PyTorch on older Intel Macs, in case we need to introduce new warnings there.