ConvBnReLU3d

class torch.ao.nn.intrinsic.qat.ConvBnReLU3d(in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=None, padding_mode='zeros', eps=1e-05, momentum=0.1, freeze_bn=False, qconfig=None)

A ConvBnReLU3d module is a module fused from Conv3d, BatchNorm3d and ReLU, attached with FakeQuantize modules for weight, used in quantization aware training.

We combined the interface of torch.nn.Conv3d and torch.nn.BatchNorm3d and torch.nn.ReLU.

Similar to torch.nn.Conv3d, with FakeQuantize modules initialized to default.

变量

weight_fake_quant – 权重的 fake quant 模块