File size: 6,808 Bytes
0c1e054
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
# Copyright (c) ByteDance, Inc. and its affiliates.
# All rights reserved.
#
# This source code is licensed under the license found in the
# LICENSE file in the root directory of this source tree.

import torch
import torch.nn as nn
from timm.models.layers import DropPath


_cur_active: torch.Tensor = None            # B1ff
# todo: try to use `gather` for speed?
def _get_active_ex_or_ii(H, W, returning_active_ex=True):
    h_repeat, w_repeat = H // _cur_active.shape[-2], W // _cur_active.shape[-1]
    active_ex = _cur_active.repeat_interleave(h_repeat, dim=2).repeat_interleave(w_repeat, dim=3)
    return active_ex if returning_active_ex else active_ex.squeeze(1).nonzero(as_tuple=True)  # ii: bi, hi, wi


def sp_conv_forward(self, x: torch.Tensor):
    x = super(type(self), self).forward(x)
    x *= _get_active_ex_or_ii(H=x.shape[2], W=x.shape[3], returning_active_ex=True)    # (BCHW) *= (B1HW), mask the output of conv
    return x


def sp_bn_forward(self, x: torch.Tensor):
    ii = _get_active_ex_or_ii(H=x.shape[2], W=x.shape[3], returning_active_ex=False)
    
    bhwc = x.permute(0, 2, 3, 1)
    nc = bhwc[ii]                               # select the features on non-masked positions to form a flatten feature `nc`
    nc = super(type(self), self).forward(nc)    # use BN1d to normalize this flatten feature `nc`
    
    bchw = torch.zeros_like(bhwc)
    bchw[ii] = nc
    bchw = bchw.permute(0, 3, 1, 2)
    return bchw


class SparseConv2d(nn.Conv2d):
    forward = sp_conv_forward   # hack: override the forward function; see `sp_conv_forward` above for more details


class SparseMaxPooling(nn.MaxPool2d):
    forward = sp_conv_forward   # hack: override the forward function; see `sp_conv_forward` above for more details


class SparseAvgPooling(nn.AvgPool2d):
    forward = sp_conv_forward   # hack: override the forward function; see `sp_conv_forward` above for more details


class SparseBatchNorm2d(nn.BatchNorm1d):
    forward = sp_bn_forward     # hack: override the forward function; see `sp_bn_forward` above for more details


class SparseSyncBatchNorm2d(nn.SyncBatchNorm):
    forward = sp_bn_forward     # hack: override the forward function; see `sp_bn_forward` above for more details


class SparseConvNeXtLayerNorm(nn.LayerNorm):
    r""" LayerNorm that supports two data formats: channels_last (default) or channels_first.
    The ordering of the dimensions in the inputs. channels_last corresponds to inputs with
    shape (batch_size, height, width, channels) while channels_first corresponds to inputs
    with shape (batch_size, channels, height, width).
    """
    
    def __init__(self, normalized_shape, eps=1e-6, data_format="channels_last", sparse=True):
        if data_format not in ["channels_last", "channels_first"]:
            raise NotImplementedError
        super().__init__(normalized_shape, eps, elementwise_affine=True)
        self.data_format = data_format
        self.sparse = sparse
    
    def forward(self, x):
        if x.ndim == 4: # BHWC or BCHW
            if self.data_format == "channels_last": # BHWC
                if self.sparse:
                    ii = _get_active_ex_or_ii(H=x.shape[1], W=x.shape[2], returning_active_ex=False)
                    nc = x[ii]
                    nc = super(SparseConvNeXtLayerNorm, self).forward(nc)
    
                    x = torch.zeros_like(x)
                    x[ii] = nc
                    return x
                else:
                    return super(SparseConvNeXtLayerNorm, self).forward(x)
            else:       # channels_first, BCHW
                if self.sparse:
                    ii = _get_active_ex_or_ii(H=x.shape[2], W=x.shape[3], returning_active_ex=False)
                    bhwc = x.permute(0, 2, 3, 1)
                    nc = bhwc[ii]
                    nc = super(SparseConvNeXtLayerNorm, self).forward(nc)
                
                    x = torch.zeros_like(bhwc)
                    x[ii] = nc
                    return x.permute(0, 3, 1, 2)
                else:
                    u = x.mean(1, keepdim=True)
                    s = (x - u).pow(2).mean(1, keepdim=True)
                    x = (x - u) / torch.sqrt(s + self.eps)
                    x = self.weight[:, None, None] * x + self.bias[:, None, None]
                    return x
        else:           # BLC or BC
            if self.sparse:
                raise NotImplementedError
            else:
                return super(SparseConvNeXtLayerNorm, self).forward(x)

    def __repr__(self):
        return super(SparseConvNeXtLayerNorm, self).__repr__()[:-1] + f', ch={self.data_format.split("_")[-1]}, sp={self.sparse})'


class SparseConvNeXtBlock(nn.Module):
    r""" ConvNeXt Block. There are two equivalent implementations:
    (1) DwConv -> LayerNorm (channels_first) -> 1x1 Conv -> GELU -> 1x1 Conv; all in (N, C, H, W)
    (2) DwConv -> Permute to (N, H, W, C); LayerNorm (channels_last) -> Linear -> GELU -> Linear; Permute back
    We use (2) as we find it slightly faster in PyTorch
    
    Args:
        dim (int): Number of input channels.
        drop_path (float): Stochastic depth rate. Default: 0.0
        layer_scale_init_value (float): Init value for Layer Scale. Default: 1e-6.
    """
    
    def __init__(self, dim, drop_path=0., layer_scale_init_value=1e-6, sparse=True, ks=7):
        super().__init__()
        self.dwconv = nn.Conv2d(dim, dim, kernel_size=ks, padding=ks//2, groups=dim)  # depthwise conv
        self.norm = SparseConvNeXtLayerNorm(dim, eps=1e-6, sparse=sparse)
        self.pwconv1 = nn.Linear(dim, 4 * dim)  # pointwise/1x1 convs, implemented with linear layers
        self.act = nn.GELU()
        self.pwconv2 = nn.Linear(4 * dim, dim)
        self.gamma = nn.Parameter(layer_scale_init_value * torch.ones((dim)),
                                  requires_grad=True) if layer_scale_init_value > 0 else None
        self.drop_path: nn.Module = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.sparse = sparse
    
    def forward(self, x):
        input = x
        x = self.dwconv(x)
        x = x.permute(0, 2, 3, 1)  # (N, C, H, W) -> (N, H, W, C)
        x = self.norm(x)
        x = self.pwconv1(x)
        x = self.act(x)            # GELU(0) == (0), so there is no need to mask x (no need to `x *= _get_active_ex_or_ii`)
        x = self.pwconv2(x)
        if self.gamma is not None:
            x = self.gamma * x
        x = x.permute(0, 3, 1, 2)  # (N, H, W, C) -> (N, C, H, W)
        
        if self.sparse:
            x *= _get_active_ex_or_ii(H=x.shape[2], W=x.shape[3], returning_active_ex=True)
        
        x = input + self.drop_path(x)
        return x
    
    def __repr__(self):
        return super(SparseConvNeXtBlock, self).__repr__()[:-1] + f', sp={self.sparse})'