Neural-Solver-Library/models/U_Net.py at main · thuml/Neural-Solver-Library · GitHub

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
import torch
import math
import torch.nn as nn
import numpy as np
import torch.nn.functional as F
from timm.models.layers import trunc_normal_
from layers.Basic import MLP
from layers.Embedding import timestep_embedding, unified_pos_embedding
from layers.UNet_Blocks import DoubleConv1D, Down1D, Up1D, OutConv1D, DoubleConv2D, Down2D, Up2D, OutConv2D, \
    DoubleConv3D, Down3D, Up3D, OutConv3D
from layers.GeoFNO_Projection import SpectralConv2d_IrregularGeo, IPHI

ConvList = [None, DoubleConv1D, DoubleConv2D, DoubleConv3D]
DownList = [None, Down1D, Down2D, Down3D]
UpList = [None, Up1D, Up2D, Up3D]
OutList = [None, OutConv1D, OutConv2D, OutConv3D]


class Model(nn.Module):
    def __init__(self, args, bilinear=True, s1=96, s2=96):
        super(Model, self).__init__()
        self.__name__ = 'U-Net'
        self.args = args
        if args.task == 'steady':
            normtype = 'bn'
        else:
            normtype = 'in'  # when conducting dynamic tasks, use instance norm for stability
        ## embedding
        if args.unified_pos and args.geotype != 'unstructured':  # only for structured mesh
            self.pos = unified_pos_embedding(args.shapelist, args.ref)
            self.preprocess = MLP(args.fun_dim + args.ref ** len(args.shapelist), args.n_hidden * 2,
                                  args.n_hidden, n_layers=0, res=False, act=args.act)
        else:
            self.preprocess = MLP(args.fun_dim + args.space_dim, args.n_hidden * 2, args.n_hidden,
                                  n_layers=0, res=False, act=args.act)
        if args.time_input:
            self.time_fc = nn.Sequential(nn.Linear(args.n_hidden, args.n_hidden), nn.SiLU(),
                                         nn.Linear(args.n_hidden, args.n_hidden))
        # geometry projection
        if self.args.geotype == 'unstructured':
            self.fftproject_in = SpectralConv2d_IrregularGeo(args.n_hidden, args.n_hidden, args.modes, args.modes,
                                                             s1, s2)
            self.fftproject_out = SpectralConv2d_IrregularGeo(args.n_hidden, args.n_hidden, args.modes, args.modes,
                                                              s1, s2)
            self.iphi = IPHI()
            patch_size = [(size + (16 - size % 16) % 16) // 16 for size in [s1, s2]]
            self.padding = [(16 - size % 16) % 16 for size in [s1, s2]]
        else:
            patch_size = [(size + (16 - size % 16) % 16) // 16 for size in args.shapelist]
            self.padding = [(16 - size % 16) % 16 for size in args.shapelist]
        # multiscale modules
        self.inc = ConvList[len(patch_size)](args.n_hidden, args.n_hidden, normtype=normtype)
        self.down1 = DownList[len(patch_size)](args.n_hidden, args.n_hidden * 2, normtype=normtype)
        self.down2 = DownList[len(patch_size)](args.n_hidden * 2, args.n_hidden * 4, normtype=normtype)
        self.down3 = DownList[len(patch_size)](args.n_hidden * 4, args.n_hidden * 8, normtype=normtype)
        factor = 2 if bilinear else 1
        self.down4 = DownList[len(patch_size)](args.n_hidden * 8, args.n_hidden * 16 // factor, normtype=normtype)
        self.up1 = UpList[len(patch_size)](args.n_hidden * 16, args.n_hidden * 8 // factor, bilinear, normtype=normtype)
        self.up2 = UpList[len(patch_size)](args.n_hidden * 8, args.n_hidden * 4 // factor, bilinear, normtype=normtype)
        self.up3 = UpList[len(patch_size)](args.n_hidden * 4, args.n_hidden * 2 // factor, bilinear, normtype=normtype)
        self.up4 = UpList[len(patch_size)](args.n_hidden * 2, args.n_hidden, bilinear, normtype=normtype)
        self.outc = OutList[len(patch_size)](args.n_hidden, args.n_hidden)
        # projectors
        self.fc1 = nn.Linear(args.n_hidden, args.n_hidden)
        self.fc2 = nn.Linear(args.n_hidden, args.out_dim)

    def multiscale(self, x):
        x1 = self.inc(x)
        x2 = self.down1(x1)
        x3 = self.down2(x2)
        x4 = self.down3(x3)
        x5 = self.down4(x4)
        x = self.up1(x5, x4)
        x = self.up2(x, x3)
        x = self.up3(x, x2)
        x = self.up4(x, x1)
        x = self.outc(x)
        return x

    def structured_geo(self, x, fx, T=None):
        B, N, _ = x.shape
        if self.args.unified_pos:
            x = self.pos.repeat(x.shape[0], 1, 1)
        if fx is not None:
            fx = torch.cat((x, fx), -1)
            fx = self.preprocess(fx)
        else:
            fx = self.preprocess(x)

        if T is not None:
            Time_emb = timestep_embedding(T, self.args.n_hidden).repeat(1, x.shape[1], 1)
            Time_emb = self.time_fc(Time_emb)
            fx = fx + Time_emb
        x = fx.permute(0, 2, 1).reshape(B, self.args.n_hidden, *self.args.shapelist)
        if not all(item == 0 for item in self.padding):
            if len(self.args.shapelist) == 2:
                x = F.pad(x, [0, self.padding[1], 0, self.padding[0]])
            elif len(self.args.shapelist) == 3:
                x = F.pad(x, [0, self.padding[2], 0, self.padding[1], 0, self.padding[0]])
        x = self.multiscale(x) ## U-Net
        if not all(item == 0 for item in self.padding):
            if len(self.args.shapelist) == 2:
                x = x[..., :-self.padding[0], :-self.padding[1]]
            elif len(self.args.shapelist) == 3:
                x = x[..., :-self.padding[0], :-self.padding[1], :-self.padding[2]]
        x = x.reshape(B, self.args.n_hidden, -1).permute(0, 2, 1)
        x = self.fc1(x)
        x = F.gelu(x)
        x = self.fc2(x)
        return x

    def unstructured_geo(self, x, fx, T=None):
        original_pos = x
        if fx is not None:
            fx = torch.cat((x, fx), -1)
            fx = self.preprocess(fx)
        else:
            fx = self.preprocess(x)

        if T is not None:
            Time_emb = timestep_embedding(T, self.args.n_hidden).repeat(1, x.shape[1], 1)
            Time_emb = self.time_fc(Time_emb)
            fx = fx + Time_emb

        x = self.fftproject_in(fx.permute(0, 2, 1), x_in=original_pos, iphi=self.iphi, code=None)
        x = self.multiscale(x) ## U-Net
        x = self.fftproject_out(x, x_out=original_pos, iphi=self.iphi, code=None).permute(0, 2, 1)
        x = self.fc1(x)
        x = F.gelu(x)
        x = self.fc2(x)
        return x

    def forward(self, x, fx, T=None, geo=None):
        if self.args.geotype == 'unstructured':
            return self.unstructured_geo(x, fx, T)
        else:
            return self.structured_geo(x, fx, T)