mqbench.nn.intrinsic.modules package

Submodules

mqbench.nn.intrinsic.modules.fused

class mqbench.nn.intrinsic.modules.fused.ConvBn2d(conv, bn)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvBnReLU2d(conv, bn, relu)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvFreezebn2d(conv, bn)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvFreezebnReLU2d(conv, bn, relu)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvReLU2d(conv, relu)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvTransposeBn2d(deconv, bn)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvTransposeBnReLU2d(deconv, bn, relu)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvTransposeFreezebn2d(deconv, bn)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvTransposeFreezebnReLU2d(deconv, bn, relu)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.ConvTransposeReLU2d(deconv, relu)[source]

Bases: _FusedModule

training: bool
class mqbench.nn.intrinsic.modules.fused.LinearBn1d(linear, bn)[source]

Bases: _FusedModule

This is a sequential container which calls the Linear and Batch Norm 1d modules. During quantization this will be replaced with the corresponding fused module.

training: bool

Module contents