brainspy#
A python package based on PyTorch and NumPy to support the study of different sorts of disordered nano-material networks for natural and energy-efficient analogue computing. Particularly, it has been applied to the concept of dopant network processing units (DNPUs), a novel and promising CMOS-compatible nano-scale tunable material with potentially very low power consumption at inference stage. The framework is focused on two material-based approaches, for training DNPUs to compute supervised learning tasks: evolution in matter and digital twin models. While the evolution in matter focuses on providing a quicker exploration of newly manufactured single units of DNPUs, the digital twin approach is used for the design and simulation of the interconnection between DNPUs, enabling to explore its scalability.
More information at: https://github.com/BraiNEdarwin/brains-py/wiki
Modules
It provides different default algorithms for brains-py, that already add several features that are particular to dopant-networks. |
|
Main package for handling all related simulations and hardware measurements of dopant-networks. |
|
A set of classes to provide support in useful tasks that are typically required when using the library. |