Artificial Superintelligence Alliance (@ASI_Alliance): FabricPC now supports convolutional architectures.

FabricPC now supports convolutional architectures. Widely used for processing spatially structured data, convolutional models underpin applications ranging from image classification to volumetric medical imaging. Predictive coding research has engaged with convolution before, but almost always through bespoke implementations tied to a single experiment. FabricPC brings convolution into a general-purpose graph framework, allowing researchers to work with convolutional architectures without building a new implementation for each experiment. The latest release introduces the library's first windowed operations, enabling CNNs and ResNets to be expressed entirely within a predictive-coding graph, in pure JAX. Visit the FabricPC GitHub repository to learn more: https://github.com/trueagi-io/FabricPC