File size: 6,643 Bytes
c46900a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
"""
Conditional U-Net Architecture for Diffusion Model
"""

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


class TimeEmbedding(nn.Module):
    def __init__(self, dim):
        super().__init__()
        self.dim = dim

    def forward(self, time):
        device = time.device
        half_dim = self.dim // 2
        embeddings = math.log(10000) / (half_dim - 1)
        embeddings = torch.exp(torch.arange(half_dim, device=device) * -embeddings)
        embeddings = time[:, None] * embeddings[None, :]
        return torch.cat([torch.sin(embeddings), torch.cos(embeddings)], dim=-1)


class LabelEmbedding(nn.Module):
    def __init__(self, label_dim, emb_dim):
        super().__init__()
        self.mlp = nn.Sequential(
            nn.Linear(label_dim, emb_dim),
            nn.SiLU(),
            nn.Linear(emb_dim, emb_dim)
        )

    def forward(self, labels):
        return self.mlp(labels)


class ResidualBlock(nn.Module):
    def __init__(self, in_channels, out_channels, time_emb_dim, dropout=0.1):
        super().__init__()
        self.conv1 = nn.Sequential(
            nn.GroupNorm(8, in_channels),
            nn.SiLU(),
            nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1)
        )
        self.time_emb = nn.Sequential(nn.SiLU(), nn.Linear(time_emb_dim, out_channels))
        self.conv2 = nn.Sequential(
            nn.GroupNorm(8, out_channels),
            nn.SiLU(),
            nn.Dropout(dropout),
            nn.Conv2d(out_channels, out_channels, kernel_size=3, padding=1)
        )
        self.shortcut = nn.Conv2d(in_channels, out_channels, kernel_size=1) if in_channels != out_channels else nn.Identity()

    def forward(self, x, emb):
        h = self.conv1(x)
        h = h + self.time_emb(emb)[:, :, None, None]
        h = self.conv2(h)
        return h + self.shortcut(x)


class AttentionBlock(nn.Module):
    def __init__(self, channels, num_heads=4):
        super().__init__()
        self.channels = channels
        self.num_heads = num_heads
        self.norm = nn.GroupNorm(8, channels)
        self.qkv = nn.Conv2d(channels, channels * 3, kernel_size=1)
        self.proj = nn.Conv2d(channels, channels, kernel_size=1)

    def forward(self, x):
        B, C, H, W = x.shape
        h = self.norm(x)
        q, k, v = self.qkv(h).chunk(3, dim=1)
        head_dim = C // self.num_heads
        q = q.view(B, self.num_heads, head_dim, H*W).transpose(2, 3)
        k = k.view(B, self.num_heads, head_dim, H*W).transpose(2, 3)
        v = v.view(B, self.num_heads, head_dim, H*W).transpose(2, 3)

        h = F.scaled_dot_product_attention(q, k, v, dropout_p=0.0)
        h = h.transpose(2, 3).reshape(B, C, H, W)
        return x + self.proj(h)


class ConditionalUNet(nn.Module):
    def __init__(self, in_channels=1, out_channels=1, label_dim=2,
                 base_channels=64, channel_multipliers=(1,2,4,8),
                 attention_levels=(2,3), dropout=0.1, time_emb_dim=256, label_emb_dim=256):
        super().__init__()
        self.label_dim = label_dim

        self.time_embedding = TimeEmbedding(time_emb_dim)
        self.time_mlp = nn.Sequential(nn.Linear(time_emb_dim, time_emb_dim*4), nn.SiLU(), nn.Linear(time_emb_dim*4, time_emb_dim))

        self.label_embedding = LabelEmbedding(label_dim, label_emb_dim)
        self.combined_emb_dim = time_emb_dim + label_emb_dim
        self.combined_mlp = nn.Sequential(nn.Linear(self.combined_emb_dim, time_emb_dim*4), nn.SiLU(), nn.Linear(time_emb_dim*4, time_emb_dim))

        self.conv_in = nn.Conv2d(in_channels, base_channels, kernel_size=3, padding=1)

        self.down_blocks = nn.ModuleList()
        channels = [base_channels]
        now_channels = base_channels

        for i, mult in enumerate(channel_multipliers):
            out_ch = base_channels * mult
            for _ in range(2):
                self.down_blocks.append(ResidualBlock(now_channels, out_ch, time_emb_dim, dropout))
                if i in attention_levels:
                    self.down_blocks.append(AttentionBlock(out_ch))
                now_channels = out_ch
                channels.append(now_channels)
            if i != len(channel_multipliers) - 1:
                self.down_blocks.append(nn.Conv2d(now_channels, now_channels, kernel_size=3, stride=2, padding=1))
                channels.append(now_channels)

        self.middle = nn.ModuleList([
            ResidualBlock(now_channels, now_channels, time_emb_dim, dropout),
            AttentionBlock(now_channels),
            ResidualBlock(now_channels, now_channels, time_emb_dim, dropout)
        ])

        self.up_blocks = nn.ModuleList()
        for i, mult in reversed(list(enumerate(channel_multipliers))):
            out_ch = base_channels * mult
            for _ in range(3):
                self.up_blocks.append(ResidualBlock(now_channels + channels.pop(), out_ch, time_emb_dim, dropout))
                if i in attention_levels:
                    self.up_blocks.append(AttentionBlock(out_ch))
                now_channels = out_ch
            if i != 0:
                self.up_blocks.append(nn.ConvTranspose2d(now_channels, now_channels, kernel_size=4, stride=2, padding=1))

        self.conv_out = nn.Sequential(
            nn.GroupNorm(8, now_channels),
            nn.SiLU(),
            nn.Conv2d(now_channels, out_channels, kernel_size=3, padding=1)
        )

    def forward(self, x, t, labels=None):
        t_emb = self.time_embedding(t)
        t_emb = self.time_mlp(t_emb)

        if labels is not None:
            label_emb = self.label_embedding(labels)
            combined = torch.cat([t_emb, label_emb], dim=-1)
            emb = self.combined_mlp(combined)
        else:
            emb = t_emb

        h = self.conv_in(x)
        skips = [h]

        for module in self.down_blocks:
            if isinstance(module, ResidualBlock):
                h = module(h, emb)
                skips.append(h)
            elif isinstance(module, AttentionBlock):
                h = module(h)
            else:
                h = module(h)
                skips.append(h)

        for module in self.middle:
            if isinstance(module, ResidualBlock):
                h = module(h, emb)
            else:
                h = module(h)

        for module in self.up_blocks:
            if isinstance(module, ResidualBlock):
                h = torch.cat([h, skips.pop()], dim=1)
                h = module(h, emb)
            elif isinstance(module, AttentionBlock):
                h = module(h)
            else:
                h = module(h)

        return self.conv_out(h)