This is an example of my first implementation of the UNET convolutional neuronal network using relatively new (version 1 released in 2018) but seems rather powerful framework (or neuronal network development and training environment) called “Neural Network Console”, created by SONY.
I should tell that this is new, highly promising system with many convenient features, including the possibility to design the network without writing any line of code, just using interactive GUI. Time permitting, I will describe this interface better.
The system is based on Miniconda distribution of Python, and includes CPU and GPU-based implementation of CNN algorithms.
Fig.1: Network editing screen