charlesmackin

It seems weight bit slicing can be implemented (within a unit cell) by simply writing a new g_converter class as seen here: https://github.com/IBM/aihwkit/blob/8cb5063eb6dc1d352e90d0786c45109de4f4ac13/src/aihwkit/inference/converter/conductance.py#L218. The CustomPairConductanceConverter class already almost does this...

@HeatPhoenix Yes, I believe it would be more appropriate (and likely simpler) to implement the synaptic bit slicing via a new ConductanceConverter class. It would be great if @maljoras could...

@HeatPhoenix Yes, you should move forward with @maljoras-sony suggestions in this case