GNERF 源码解读（二）

生成器架构

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
GNeRF(
  (nerf): NeRF(
    (embedding_xyz): Embedding()
    (embedding_dir): Embedding()
    (xyz_encoding_1): Sequential(
      (0): Linear(in_features=63, out_features=360, bias=True)
      (1): ReLU(inplace=True)
    )
    (xyz_encoding_2): Sequential(
      (0): Linear(in_features=360, out_features=360, bias=True)
      (1): ReLU(inplace=True)
    )
    (xyz_encoding_3): Sequential(
      (0): Linear(in_features=360, out_features=360, bias=True)
      (1): ReLU(inplace=True)
    )
    (xyz_encoding_4): Sequential(
      (0): Linear(in_features=360, out_features=360, bias=True)
      (1): ReLU(inplace=True)
    )
    (xyz_encoding_5): Sequential(
      (0): Linear(in_features=423, out_features=360, bias=True)
      (1): ReLU(inplace=True)
    )
    (xyz_encoding_6): Sequential(
      (0): Linear(in_features=360, out_features=360, bias=True)
      (1): ReLU(inplace=True)
    )
    (xyz_encoding_7): Sequential(
      (0): Linear(in_features=360, out_features=360, bias=True)
      (1): ReLU(inplace=True)
    )
    (xyz_encoding_8): Sequential(
      (0): Linear(in_features=360, out_features=360, bias=True)
      (1): ReLU(inplace=True)
    )
    (xyz_encoding_final): Linear(in_features=360, out_features=360, bias=True)
    (sigma): Linear(in_features=360, out_features=1, bias=True)
    (rgb): Sequential(
      (0): Linear(in_features=423, out_features=180, bias=True)
      (1): ReLU(inplace=True)
      (2): Linear(in_features=180, out_features=3, bias=True)
      (3): Sigmoid()
    )
  )
)

辨别器架构

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Discriminator(
  (embedding_scale): Embedding()
  (final): Sequential(
    (0): LeakyReLU(negative_slope=0.2)
    (1): Conv2d(73, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
    (2): LeakyReLU(negative_slope=0.2, inplace=True)
    (3): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
    (4): LeakyReLU(negative_slope=0.2, inplace=True)
    (5): Conv2d(64, 1, kernel_size=(1, 1), stride=(1, 1), bias=False)
  )
  (main): Sequential(
    (0): Conv2d(3, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (1): InstanceNorm2d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (2): LeakyReLU(negative_slope=0.2, inplace=True)
    (3): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (4): InstanceNorm2d(512, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (5): LeakyReLU(negative_slope=0.2, inplace=True)
    (6): Conv2d(512, 64, kernel_size=(4, 4), stride=(1, 1), bias=False)
  )
)
源码解读
#3D GAN
GNERF 源码解读（二）
http://seulqxq.top/posts/39622/
作者
SeulQxQ
发布于
2024年2月27日
许可协议