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
|
from einops.layers.torch import Rearrange
from torch import Tensor, nn
from .transformer.decoder import Decoder
from .transformer.embedding.sincos import sincos_2d
from .transformer.embedding.token import TokenEmbedding
from .transformer.encoder import Encoder
class Convformer(nn.Module):
def __init__(
self,
image_height: int,
image_width: int,
patch_height: int,
patch_width: int,
dim: int,
num_classes: int,
encoder: Encoder,
decoder: Decoder,
token_embedding: TokenEmbedding,
tie_embeddings: bool,
pad_index: int,
channels: int = 1,
) -> None:
super().__init__()
patch_dim = patch_height * patch_width * channels
self.to_patch_embedding = nn.Sequential(
Rearrange(
"b c (h ph) (w pw) -> b (h w) (ph pw c)",
ph=patch_height,
pw=patch_width,
),
nn.LayerNorm(patch_dim),
nn.Linear(patch_dim, dim),
nn.LayerNorm(dim),
)
self.patch_embedding = sincos_2d(
h=image_height // patch_height, w=image_width // patch_width, dim=dim
)
self.token_embedding = token_embedding
self.to_logits = (
nn.Linear(dim, num_classes)
if not tie_embeddings
else lambda t: t @ self.token_embedding.to_embedding.weight.t()
)
self.encoder = encoder
self.decoder = decoder
self.pad_index = pad_index
def encode(self, images: Tensor) -> Tensor:
x = self.encoder(images)
x = self.to_patch_embedding(x)
x = x + self.patch_embedding.to(images.device, dtype=images.dtype)
return x
def decode(self, text: Tensor, img_features: Tensor) -> Tensor:
text = text.long()
mask = text != self.pad_index
tokens = self.token_embedding(text)
output = self.decoder(tokens, context=img_features, mask=mask)
return self.to_logits(output)
def forward(
self,
img: Tensor,
text: Tensor,
) -> Tensor:
"""Applies decoder block on input signals."""
img_features = self.encode(img)
logits = self.decode(text, img_features)
return logits # [B, N, C]
|
