aihwkit.inference.compensation.base module¶

Base drift compensation for inference.

class aihwkit.inference.compensation.base.BaseDriftCompensation[source]¶

Bases: object

Base class for drift compensations.

apply(forward_output, ref_value)[source]¶

Read out the current value from the output of the forward pass and returns the drift compensation alpha scale.

get_readout_tensor(in_size)[source]¶

Return the read-out tensor.

Called once during init_baseline().

init_baseline(forward_output)[source]¶

Initialize the base line for applying the compensation.

Uses a all one tensor for read_out.

Parameters: forward_output – forward output of the read out vector to compensate
Returns: reference tensor readout
Return type: Tuple[torch.Tensor, torch.Tensor]

readout(out_tensor)[source]¶

Implement the read out math.