(Customized) Big Transfer (BiT) Models from google-research/big_transfer

bit_models.py

# Adapted from: https://github.com/google-research/big_transfer/blob/6c83d6459e009fa89d84c1e904611e9b162e6eff/bit_pytorch/models.py
# Copyright 2020 Google LLC
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

# Lint as: python3
"""Bottleneck ResNet v2 with GroupNorm and Weight Standardization."""

from collections import OrderedDict  # pylint: disable=g-importing-member

import torch
import torch.nn as nn
import torch.nn.functional as F


class StdConv2d(nn.Conv2d):
    def forward(self, x):
        w = self.weight
        v, m = torch.var_mean(w, dim=[1, 2, 3], keepdim=True, unbiased=False)
        w = (w - m) / torch.sqrt(v + 1e-10)
        return F.conv2d(x, w, self.bias, self.stride, self.padding,
                        self.dilation, self.groups)


def conv3x3(cin, cout, stride=1, groups=1, bias=False):
    return StdConv2d(cin, cout, kernel_size=3, stride=stride,
                     padding=1, bias=bias, groups=groups)


def conv1x1(cin, cout, stride=1, bias=False):
    return StdConv2d(cin, cout, kernel_size=1, stride=stride,
                     padding=0, bias=bias)


def tf2th(conv_weights):
    """Possibly convert HWIO to OIHW."""
    if conv_weights.ndim == 4:
        conv_weights = conv_weights.transpose([3, 2, 0, 1])
    return torch.from_numpy(conv_weights)


class PreActBottleneck(nn.Module):
    """Pre-activation (v2) bottleneck block.

    Follows the implementation of "Identity Mappings in Deep Residual Networks":
    https://github.com/KaimingHe/resnet-1k-layers/blob/master/resnet-pre-act.lua

    Except it puts the stride on 3x3 conv when available.
    """

    def __init__(self, cin, cout=None, cmid=None, stride=1):
        super().__init__()
        cout = cout or cin
        cmid = cmid or cout//4

        self.gn1 = nn.GroupNorm(32, cin)
        self.conv1 = conv1x1(cin, cmid)
        self.gn2 = nn.GroupNorm(32, cmid)
        # Original code has it on conv1!!
        self.conv2 = conv3x3(cmid, cmid, stride)
        self.gn3 = nn.GroupNorm(32, cmid)
        self.conv3 = conv1x1(cmid, cout)
        self.relu = nn.ReLU(inplace=True)

        if (stride != 1 or cin != cout):
            # Projection also with pre-activation according to paper.
            self.downsample = conv1x1(cin, cout, stride)

    def forward(self, x):
        out = self.relu(self.gn1(x))

        # Residual branch
        residual = x
        if hasattr(self, 'downsample'):
            residual = self.downsample(out)

        # Unit's branch
        out = self.conv1(out)
        out = self.conv2(self.relu(self.gn2(out)))
        out = self.conv3(self.relu(self.gn3(out)))

        return out + residual

    def load_from(self, weights, prefix=''):
        convname = 'standardized_conv2d'
        with torch.no_grad():
            self.conv1.weight.copy_(
                tf2th(weights[f'{prefix}a/{convname}/kernel']))
            self.conv2.weight.copy_(
                tf2th(weights[f'{prefix}b/{convname}/kernel']))
            self.conv3.weight.copy_(
                tf2th(weights[f'{prefix}c/{convname}/kernel']))
            self.gn1.weight.copy_(
                tf2th(weights[f'{prefix}a/group_norm/gamma']))
            self.gn2.weight.copy_(
                tf2th(weights[f'{prefix}b/group_norm/gamma']))
            self.gn3.weight.copy_(
                tf2th(weights[f'{prefix}c/group_norm/gamma']))
            self.gn1.bias.copy_(tf2th(weights[f'{prefix}a/group_norm/beta']))
            self.gn2.bias.copy_(tf2th(weights[f'{prefix}b/group_norm/beta']))
            self.gn3.bias.copy_(tf2th(weights[f'{prefix}c/group_norm/beta']))
            if hasattr(self, 'downsample'):
                w = weights[f'{prefix}a/proj/{convname}/kernel']
                self.downsample.weight.copy_(tf2th(w))


class ResNetV2(nn.Module):
    """Implementation of Pre-activation (v2) ResNet mode."""

    def __init__(self, block_units, width_factor, head_size=21843, zero_head=False):
        """ResNetV2

        Parameters
        ----------
        block_units : Size of the blocks
        width_factor : width scaling factor
        head_size : int, optional
            Set -1 to remove final projection, by default 21843
        zero_head : bool, optional
            Set the weights of the head to zeros, by default False
        """
        super().__init__()
        self.head_size = head_size
        self.zero_head = zero_head
        self.width_factor = width_factor
        wf = width_factor  # shortcut 'cause we'll use it a lot.

        # The following will be unreadable if we split lines.
        # pylint: disable=line-too-long
        self.root = nn.Sequential(OrderedDict([
            ('conv', StdConv2d(3, 64*wf, kernel_size=7,
                               stride=2, padding=3, bias=False)),
            ('pad', nn.ConstantPad2d(1, 0.)),
            ('pool', nn.MaxPool2d(kernel_size=3, stride=2, padding=0)),
            # The following is subtly not the same!
            # ('pool', nn.MaxPool2d(kernel_size=3, stride=2, padding=1)),
        ]))

        self.body = nn.Sequential(OrderedDict([
            ('block1', nn.Sequential(OrderedDict(
                [('unit01', PreActBottleneck(cin=64*wf, cout=256*wf, cmid=64*wf))] +
                [(f'unit{i:02d}', PreActBottleneck(cin=256*wf, cout=256*wf, cmid=64*wf))
                 for i in range(2, block_units[0] + 1)],
            ))),
            ('block2', nn.Sequential(OrderedDict(
                [('unit01', PreActBottleneck(cin=256*wf, cout=512*wf, cmid=128*wf, stride=2))] +
                [(f'unit{i:02d}', PreActBottleneck(cin=512*wf, cout=512 *
                                                   wf, cmid=128*wf)) for i in range(2, block_units[1] + 1)],
            ))),
            ('block3', nn.Sequential(OrderedDict(
                [('unit01', PreActBottleneck(cin=512*wf, cout=1024*wf, cmid=256*wf, stride=2))] +
                [(f'unit{i:02d}', PreActBottleneck(cin=1024*wf, cout=1024 *
                                                   wf, cmid=256*wf)) for i in range(2, block_units[2] + 1)],
            ))),
            ('block4', nn.Sequential(OrderedDict(
                [('unit01', PreActBottleneck(cin=1024*wf, cout=2048*wf, cmid=512*wf, stride=2))] +
                [(f'unit{i:02d}', PreActBottleneck(cin=2048*wf, cout=2048 *
                                                   wf, cmid=512*wf)) for i in range(2, block_units[3] + 1)],
            ))),
        ]))
        # pylint: enable=line-too-long
        if head_size <= 0:
            self.head = nn.Sequential(OrderedDict([
                ('gn', nn.GroupNorm(32, 2048*wf)),
                ('relu', nn.ReLU(inplace=True)),
                ('avg', nn.AdaptiveAvgPool2d(output_size=1)),
            ]))
        else:
            self.head = nn.Sequential(OrderedDict([
                ('gn', nn.GroupNorm(32, 2048*wf)),
                ('relu', nn.ReLU(inplace=True)),
                ('avg', nn.AdaptiveAvgPool2d(output_size=1)),
                ('conv', nn.Conv2d(2048*wf, head_size, kernel_size=1, bias=True)),
            ]))

    def forward(self, x):
        x = self.head(self.body(self.root(x)))
        assert x.shape[-2:] == (1, 1)  # We should have no spatial shape left.
        return x[..., 0, 0]

    def load_from(self, weights, prefix='resnet/'):
        with torch.no_grad():
            self.root.conv.weight.copy_(tf2th(
                weights[f'{prefix}root_block/standardized_conv2d/kernel']))  # pylint: disable=line-too-long
            self.head.gn.weight.copy_(
                tf2th(weights[f'{prefix}group_norm/gamma']))
            self.head.gn.bias.copy_(tf2th(weights[f'{prefix}group_norm/beta']))
            if self.zero_head:
                nn.init.zeros_(self.head.conv.weight)
                nn.init.zeros_(self.head.conv.bias)
            elif self.head_size > 0:
                self.head.conv.weight.copy_(
                    tf2th(weights[f'{prefix}head/conv2d/kernel']))  # pylint: disable=line-too-long
                self.head.conv.bias.copy_(
                    tf2th(weights[f'{prefix}head/conv2d/bias']))

            for bname, block in self.body.named_children():
                for uname, unit in block.named_children():
                    unit.load_from(weights, prefix=f'{prefix}{bname}/{uname}/')


KNOWN_MODELS = OrderedDict([
    ('BiT-M-R50x1', lambda *a, **kw: ResNetV2([3, 4, 6, 3], 1, *a, **kw)),
    ('BiT-M-R50x3', lambda *a, **kw: ResNetV2([3, 4, 6, 3], 3, *a, **kw)),
    ('BiT-M-R101x1', lambda *a, **kw: ResNetV2([3, 4, 23, 3], 1, *a, **kw)),
    ('BiT-M-R101x3', lambda *a, **kw: ResNetV2([3, 4, 23, 3], 3, *a, **kw)),
    ('BiT-M-R152x2', lambda *a, **kw: ResNetV2([3, 8, 36, 3], 2, *a, **kw)),
    ('BiT-M-R152x4', lambda *a, **kw: ResNetV2([3, 8, 36, 3], 4, *a, **kw)),
    ('BiT-S-R50x1', lambda *a, **kw: ResNetV2([3, 4, 6, 3], 1, *a, **kw)),
    ('BiT-S-R50x3', lambda *a, **kw: ResNetV2([3, 4, 6, 3], 3, *a, **kw)),
    ('BiT-S-R101x1', lambda *a, **kw: ResNetV2([3, 4, 23, 3], 1, *a, **kw)),
    ('BiT-S-R101x3', lambda *a, **kw: ResNetV2([3, 4, 23, 3], 3, *a, **kw)),
    ('BiT-S-R152x2', lambda *a, **kw: ResNetV2([3, 8, 36, 3], 2, *a, **kw)),
    ('BiT-S-R152x4', lambda *a, **kw: ResNetV2([3, 8, 36, 3], 4, *a, **kw)),
])