wide_resnet_mge.py

#!/usr/bin/env python
# coding: utf-8

import megengine as mge
import megengine.functional as F
import megengine.module as M


def conv3x3(in_planes: int, out_planes: int, stride: int = 1, groups: int = 1, dilation: int = 1):
    """3x3 convolution with padding"""
    return M.Conv2d(
        in_planes,
        out_planes,
        kernel_size=3,
        stride=stride,
        padding=dilation,
        groups=groups,
        bias=False,
        dilation=dilation,
    )


def conv1x1(in_planes: int, out_planes: int, stride: int = 1):
    """1x1 convolution"""
    return M.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, bias=False)


class Bottleneck(M.Module):
    expansion: int = 4

    def __init__(
            self,
            inplanes: int,
            planes: int,
            stride: int = 1,
            downsample=None,
            groups: int = 1,
            base_width: int = 64,
            dilation: int = 1,
            norm_layer=None,
    ) -> None:
        super().__init__()
        if norm_layer is None:
            norm_layer = M.BatchNorm2d
        width = int(planes * (base_width / 64.0)) * groups
        # Both self.conv2 and self.downsample layers downsample the input when stride != 1
        self.conv1 = conv1x1(inplanes, width)
        self.bn1 = norm_layer(width)
        self.conv2 = conv3x3(width, width, stride, groups, dilation)
        self.bn2 = norm_layer(width)
        self.conv3 = conv1x1(width, planes * self.expansion)
        self.bn3 = norm_layer(planes * self.expansion)
        self.relu = M.ReLU()
        self.downsample = downsample
        self.stride = stride

    def forward(self, x):
        identity = x

        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        out = self.conv2(out)
        out = self.bn2(out)
        out = self.relu(out)

        out = self.conv3(out)
        out = self.bn3(out)

        if self.downsample is not None:
            identity = self.downsample(x)

        out += identity
        out = self.relu(out)

        return out


class WideResNet(M.Module):
    def __init__(
            self,
            block,
            layers,
            num_classes: int = 1000,
            zero_init_residual: bool = False,
            groups: int = 1,
            width_per_group: int = 64,
            replace_stride_with_dilation=None,
            norm_layer=None,
    ) -> None:
        super().__init__()
        if norm_layer is None:
            norm_layer = M.BatchNorm2d
        self._norm_layer = norm_layer

        self.inplanes = 64
        self.dilation = 1
        if replace_stride_with_dilation is None:
            # each element in the tuple indicates if we should replace
            # the 2x2 stride with a dilated convolution instead
            replace_stride_with_dilation = [False, False, False]
        if len(replace_stride_with_dilation) != 3:
            raise ValueError(
                "replace_stride_with_dilation should be None "
                f"or a 3-element tuple, got {replace_stride_with_dilation}"
            )
        self.groups = groups
        self.base_width = width_per_group
        self.conv1 = M.Conv2d(3, self.inplanes, kernel_size=7, stride=2, padding=3, bias=False)
        self.bn1 = norm_layer(self.inplanes)
        self.relu = M.ReLU()
        self.maxpool = M.MaxPool2d(kernel_size=3, stride=2, padding=1)
        self.layer1 = self._make_layer(block, 64, layers[0])
        self.layer2 = self._make_layer(block, 128, layers[1], stride=2, dilate=replace_stride_with_dilation[0])
        self.layer3 = self._make_layer(block, 256, layers[2], stride=2, dilate=replace_stride_with_dilation[1])
        self.layer4 = self._make_layer(block, 512, layers[3], stride=2, dilate=replace_stride_with_dilation[2])
        self.avgpool = M.AdaptiveAvgPool2d((1, 1))
        self.fc = M.Linear(512 * block.expansion, num_classes)

    def _make_layer(
            self,
            block,
            planes: int,
            blocks: int,
            stride: int = 1,
            dilate: bool = False,
    ) -> M.Sequential:
        norm_layer = self._norm_layer
        downsample = None
        previous_dilation = self.dilation
        if dilate:
            self.dilation *= stride
            stride = 1
        if stride != 1 or self.inplanes != planes * block.expansion:
            downsample = M.Sequential(
                conv1x1(self.inplanes, planes * block.expansion, stride),
                norm_layer(planes * block.expansion),
            )

        layers = []
        layers.append(
            block(
                self.inplanes, planes, stride, downsample, self.groups, self.base_width, previous_dilation, norm_layer
            )
        )
        self.inplanes = planes * block.expansion
        for _ in range(1, blocks):
            layers.append(
                block(
                    self.inplanes,
                    planes,
                    groups=self.groups,
                    base_width=self.base_width,
                    dilation=self.dilation,
                    norm_layer=norm_layer,
                )
            )

        return M.Sequential(*layers)

    def _forward_impl(self, x):
        # See note [TorchScript super()]
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)

        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)

        x = self.avgpool(x)
        x = F.flatten(x, 1, -1)

        x = self.fc(x)

        return x

    def forward(self, x):
        return self._forward_impl(x)


import megengine.hub as hub


@hub.pretrained(
    'https://studio.brainpp.com/api/v1/activities/3/missions/89/files/24cb5aac-c359-44ae-97f2-e7b5a0c1fc5c'
)
def wide_resnet50_2():
    return WideResNet(Bottleneck, [3, 4, 6, 3], width_per_group=128)


if __name__ == '__main__':
    import torch

    torch_checkpoint = torch.load('./wide_resnet50_2-9ba9bcbe.pth')
    # mge_model = WideResNet(Bottleneck, [3, 4, 6, 3])
    mge_model = wide_resnet50_2()

    state_dict = mge_model.state_dict()
    from convert_torch2mge import convert2mge

    new_state = convert2mge(torch_checkpoint, state_dict)
    mge.save(new_state, './wide_resnet50_2.mge')
    # print('sssssssss')
    state_dict = mge.load("./wide_resnet50_2.mge")
    # mge_model = wide_resnet50_2()
    mge_model.load_state_dict(state_dict)
    print(mge_model)