Experimenting with the forward-foward algorithm based on https://keras.io/examples/vision/forwardforward/, implementing Geoffrey Hinton's paper https://arxiv.org/abs/2212.13345
Having built a working FF MLP, the next step is to build an image classifier. The first attempt is to use a forward-forward network as the classifier on top of a pretrained CNN, such as Mobilenet_v2.
Currently, the training of the model appears to be functioning.