quark.torch.quantization.nn.modules.quantize_conv_bn_fused

Module Contents

Classes

Functions

class quark.torch.quantization.nn.modules.quantize_conv_bn_fused.QuantizedConvBatchNorm2d(in_channels: int, out_channels: int, kernel_size: torch.nn.common_types._size_2_t, stride: torch.nn.common_types._size_2_t = 1, padding: torch.nn.common_types._size_2_t = 0, dilation: torch.nn.common_types._size_2_t = 1, groups: int = 1, bias: bool = True, padding_mode: str = 'zeros', eps: float = 1e-05, momentum: float = 0.1, freeze_bn_stats: bool = False, quant_config: Optional[quark.torch.quantization.config.config.QuantizationConfig] = QuantizationConfig())

A QuantizedConvBatchNorm2d module is a module fused from

Conv2d and BatchNorm2d attached with FakeQuantizer modules for weight and batchnorm stuffs used in quantization aware training.

We combined the interface of :class:`torch.nn.Conv2d` and :class:`torch.nn.BatchNorm2d`.

Implementation details: https://arxiv.org/pdf/1806.08342.pdf section 3.2.2

Similar to :class:`torch.nn.Conv2d`, with FakeQuantizer modules initialized to default.

#