Source code for bitorch_engine.layers.qlinear.binary

from .layer import BinaryLinearBase, BinaryLinearParameter
import torch.cuda



[docs]
def get_best_binary_implementation():
    if torch.cuda.is_available():
        from .cuda import BinaryLinearCuda
        return BinaryLinearCuda
    else:
        from .cpp import BinaryLinearCPP
        return BinaryLinearCPP



BinaryLinear = get_best_binary_implementation()