Spaces:
Running
Running
Update discriminatorModel.py
Browse files- discriminatorModel.py +15 -15
discriminatorModel.py
CHANGED
@@ -2,6 +2,21 @@ import torch
|
|
2 |
import torch.nn as nn
|
3 |
|
4 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
# The Discriminator model
|
6 |
class Discriminator(nn.Module):
|
7 |
def __init__(self, channels, embed_dim=1024, embed_out_dim=128):
|
@@ -23,21 +38,6 @@ class Discriminator(nn.Module):
|
|
23 |
nn.Conv2d(512 + self.embed_out_dim, 1, 4, 1, 0, bias=False), nn.Sigmoid()
|
24 |
)
|
25 |
|
26 |
-
class Embedding(nn.Module):
|
27 |
-
def __init__(self, size_in, size_out):
|
28 |
-
super(Embedding, self).__init__()
|
29 |
-
self.text_embedding = nn.Sequential(
|
30 |
-
nn.Linear(size_in, size_out),
|
31 |
-
nn.BatchNorm1d(1),
|
32 |
-
nn.LeakyReLU(0.2, inplace=True)
|
33 |
-
)
|
34 |
-
|
35 |
-
def forward(self, x, text):
|
36 |
-
embed_out = self.text_embedding(text)
|
37 |
-
embed_out_resize = embed_out.repeat(4, 1, 4, 1).permute(1, 3, 0, 2) # Resize to match the discriminator input size
|
38 |
-
out = torch.cat([x, embed_out_resize], 1) # Concatenate text embedding with the input feature map
|
39 |
-
return out
|
40 |
-
|
41 |
def _create_layer(self, size_in, size_out, kernel_size=4, stride=2, padding=1, normalize=True):
|
42 |
layers = [nn.Conv2d(size_in, size_out, kernel_size=kernel_size, stride=stride, padding=padding)]
|
43 |
if normalize:
|
|
|
2 |
import torch.nn as nn
|
3 |
|
4 |
|
5 |
+
class Embedding(nn.Module):
|
6 |
+
def __init__(self, size_in, size_out):
|
7 |
+
super(Embedding, self).__init__()
|
8 |
+
self.text_embedding = nn.Sequential(
|
9 |
+
nn.Linear(size_in, size_out),
|
10 |
+
nn.BatchNorm1d(1),
|
11 |
+
nn.LeakyReLU(0.2, inplace=True)
|
12 |
+
)
|
13 |
+
|
14 |
+
def forward(self, x, text):
|
15 |
+
embed_out = self.text_embedding(text)
|
16 |
+
embed_out_resize = embed_out.repeat(4, 1, 4, 1).permute(1, 3, 0, 2) # Resize to match the discriminator input size
|
17 |
+
out = torch.cat([x, embed_out_resize], 1) # Concatenate text embedding with the input feature map
|
18 |
+
return out
|
19 |
+
|
20 |
# The Discriminator model
|
21 |
class Discriminator(nn.Module):
|
22 |
def __init__(self, channels, embed_dim=1024, embed_out_dim=128):
|
|
|
38 |
nn.Conv2d(512 + self.embed_out_dim, 1, 4, 1, 0, bias=False), nn.Sigmoid()
|
39 |
)
|
40 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
41 |
def _create_layer(self, size_in, size_out, kernel_size=4, stride=2, padding=1, normalize=True):
|
42 |
layers = [nn.Conv2d(size_in, size_out, kernel_size=kernel_size, stride=stride, padding=padding)]
|
43 |
if normalize:
|