File size: 5,768 Bytes
02c5426 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 |
import torch
import torch.nn as nn
import torch.nn.functional as F
from einops import rearrange, repeat
MIN_NUM_PATCHES = 16
"""
This is a new remote sensing super-resolution method based on the prevalent transformer
ref:
https://github.com/lucidrains/vit-pytorch/blob/main/vit_pytorch/vit_pytorch.py
"""
class Residual(nn.Module):
def __init__(self, fn):
super().__init__()
self.fn = fn
def forward(self, x, **kwargs):
return self.fn(x, **kwargs) + x
class Residual2(nn.Module):
def __init__(self, fn):
super().__init__()
self.fn = fn
def forward(self, x, m=None, **kwargs):
return self.fn(x, m, **kwargs) + x
class PreNorm(nn.Module):
def __init__(self, dim, fn):
super().__init__()
self.norm = nn.LayerNorm(dim)
self.fn = fn
def forward(self, x, **kwargs):
return self.fn(self.norm(x), **kwargs)
class PreNorm2(nn.Module):
def __init__(self, dim, fn):
super().__init__()
self.norm = nn.LayerNorm(dim)
self.fn = fn
def forward(self, x, m=None, **kwargs):
x = self.norm(x)
if m is not None: m = self.norm(m)
return self.fn(x, m, **kwargs)
class FeedForward(nn.Module):
def __init__(self, dim, hidden_dim, dropout = 0.):
super().__init__()
self.net = nn.Sequential(
nn.Linear(dim, hidden_dim),
nn.GELU(),
nn.Dropout(dropout),
nn.Linear(hidden_dim, dim),
nn.Dropout(dropout)
)
def forward(self, x):
return self.net(x)
class Attention(nn.Module):
def __init__(self, dim, heads = 8, dim_head = 64, dropout = 0.):
super().__init__()
inner_dim = dim_head * heads
self.heads = heads
self.scale = dim ** -0.5
self.to_qkv = nn.Linear(dim, inner_dim * 3, bias = False)
self.to_out = nn.Sequential(
nn.Linear(inner_dim, dim),
nn.Dropout(dropout)
)
def forward(self, x, mask = None):
b, n, _, h = *x.shape, self.heads
qkv = self.to_qkv(x).chunk(3, dim = -1)
q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> b h n d', h = h), qkv)
dots = torch.einsum('bhid,bhjd->bhij', q, k) * self.scale
mask_value = -torch.finfo(dots.dtype).max
if mask is not None:
mask = F.pad(mask.flatten(1), (1, 0), value = True)
assert mask.shape[-1] == dots.shape[-1], 'mask has incorrect dimensions'
mask = mask[:, None, :] * mask[:, :, None]
dots.masked_fill_(~mask, mask_value)
del mask
attn = dots.softmax(dim=-1)
out = torch.einsum('bhij,bhjd->bhid', attn, v)
out = rearrange(out, 'b h n d -> b n (h d)')
out = self.to_out(out)
return out
class MixedAttention(nn.Module):
def __init__(self, dim, heads=8, dim_head=64, dropout=0.):
super().__init__()
inner_dim = dim_head * heads
self.heads = heads
self.scale = dim ** -0.5
self.to_q = nn.Linear(dim, inner_dim, bias=False)
self.to_k = nn.Linear(dim, inner_dim, bias=False)
self.to_v = nn.Linear(dim, inner_dim, bias=False)
self.to_out = nn.Sequential(
nn.Linear(inner_dim, dim),
nn.Dropout(dropout)
)
def forward(self, x, m, mask=None):
b, n, _, h = *x.shape, self.heads
q = self.to_q(x)
k = self.to_k(m)
v = self.to_v(m)
q = rearrange(q, 'b n (h d) -> b h n d', h=h)
k = rearrange(k, 'b n (h d) -> b h n d', h=h)
v = rearrange(v, 'b n (h d) -> b h n d', h=h)
dots = torch.einsum('bhid,bhjd->bhij', q, k) * self.scale
mask_value = -torch.finfo(dots.dtype).max
if mask is not None:
mask = F.pad(mask.flatten(1), (1, 0), value = True)
assert mask.shape[-1] == dots.shape[-1], 'mask has incorrect dimensions'
mask = mask[:, None, :] * mask[:, :, None]
dots.masked_fill_(~mask, mask_value)
del mask
attn = dots.softmax(dim=-1)
out = torch.einsum('bhij,bhjd->bhid', attn, v)
out = rearrange(out, 'b h n d -> b n (h d)')
out = self.to_out(out)
return out
class TransformerEncoder(nn.Module):
def __init__(self, dim, depth, heads, dim_head, mlp_dim, dropout):
super().__init__()
self.layers = nn.ModuleList([])
for _ in range(depth):
self.layers.append(nn.ModuleList([
Residual(PreNorm(dim, Attention(dim, heads=heads, dim_head=dim_head, dropout=dropout))),
Residual(PreNorm(dim, FeedForward(dim, mlp_dim, dropout=dropout)))
]))
def forward(self, x, mask=None):
for attn, ff in self.layers:
x = attn(x, mask=mask)
x = ff(x)
return x
class TransformerDecoder(nn.Module):
def __init__(self, dim, depth, heads, dim_head, mlp_dim, dropout):
super().__init__()
self.layers = nn.ModuleList([])
for _ in range(depth):
self.layers.append(nn.ModuleList([
Residual(PreNorm(dim, Attention(dim, heads=heads, dim_head=dim_head, dropout=dropout))),
Residual2(PreNorm2(dim, MixedAttention(dim, heads=heads, dim_head=dim_head, dropout=dropout))),
Residual(PreNorm(dim, FeedForward(dim, mlp_dim, dropout=dropout)))
]))
def with_pos_embed(self, tensor, pos=None):
return tensor if pos is None else tensor + pos
def forward(self, x, m, mask=None):
for attn1, attn2, ff in self.layers:
x = attn1(x, mask=mask)
x = attn2(x, m, mask=mask)
x = ff(x)
return x |