|
|
|
|
|
r""" |
|
|
This file is in the process of migration to `torch/ao/quantization`, and |
|
|
is kept here for compatibility while the migration process is ongoing. |
|
|
If you are adding a new entry/functionality, please, add it to the |
|
|
appropriate files under `torch/ao/quantization/fx/`, while adding an import statement |
|
|
here. |
|
|
""" |
|
|
from torch.ao.quantization.fx.quantization_patterns import ( |
|
|
QuantizeHandler, |
|
|
BinaryOpQuantizeHandler, |
|
|
CatQuantizeHandler, |
|
|
ConvReluQuantizeHandler, |
|
|
LinearReLUQuantizeHandler, |
|
|
BatchNormQuantizeHandler, |
|
|
EmbeddingQuantizeHandler, |
|
|
RNNDynamicQuantizeHandler, |
|
|
DefaultNodeQuantizeHandler, |
|
|
FixedQParamsOpQuantizeHandler, |
|
|
CopyNodeQuantizeHandler, |
|
|
CustomModuleQuantizeHandler, |
|
|
GeneralTensorShapeOpQuantizeHandler, |
|
|
StandaloneModuleQuantizeHandler |
|
|
) |
|
|
|
|
|
QuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
BinaryOpQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
CatQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
ConvReluQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
LinearReLUQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
BatchNormQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
EmbeddingQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
RNNDynamicQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
DefaultNodeQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
FixedQParamsOpQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
CopyNodeQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
CustomModuleQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
GeneralTensorShapeOpQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
StandaloneModuleQuantizeHandler.__module__ = "torch.quantization.fx.quantization_patterns" |
|
|
|
