In the paper, the author wrote as follows: "In Fig. 2(e), the first-layer feature

Why does the 7-th layer of stylegan2 have the resolution as same as 32x32. about styleganex HOT 1 OPEN

DongMyongJong commented on May 30, 2024

Why does the 7-th layer of stylegan2 have the resolution as same as 32x32.

from styleganex.

Comments (1)

williamyang1991 commented on May 30, 2024

StyleGANEX/models/stylegan2/model.py

Lines 456 to 469 in e593d16

    
           self.log_size = int(math.log(size, 2)) 
        
           self.num_layers = (self.log_size - 2) * 2 + 1 
        
           self.convs = nn.ModuleList() 
        
           self.upsamples = nn.ModuleList() 
        
           self.to_rgbs = nn.ModuleList() 
        
           self.noises = nn.Module() 
        
           in_channel = self.channels[4] 
        
           for layer_idx in range(self.num_layers): 
        
               res = (layer_idx + 5) // 2 
        
               shape = [1, 1, 2 ** res, 2 ** res] 
        
               self.noises.register_buffer(f'noise_{layer_idx}', torch.randn(*shape))

begin from 0, the 6th layer is 32x32
begin from 1, the 7th layer is 32x32

from styleganex.

Related Issues (20)

Traceback error HOT 1
Problem with launching Gradio on WIndows HOT 3
Face attribute editing HOT 1
Face Restoration HOT 1
inversion HOT 5
关于输入输出分辨率不一致问题，如何才能保证输入输出分辨率一致？ HOT 2
about the padding calculation when use dilation HOT 5
Some confusion HOT 4
Image edit HOT 2
Video output size HOT 1
Edit Vector HOT 14
when using huggingface "Image Face Editing" strange image are generated. The texture looks like "leather"
Ajust Age HOT 2
How to convert the model to onnx? HOT 2
SyntaxError: 'return' outside function` HOT 3
How did you find the editing_w styles for style transfer? HOT 1
Video face editing HOT 2
pretrain weight stylegan2-ffhq-config-f.pt HOT 1
Video editing output size HOT 9

Why does the 7-th layer of stylegan2 have the resolution as same as 32x32. about styleganex HOT 1 OPEN

Comments (1)

Related Issues (20)

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent

	self.log_size = int(math.log(size, 2))
	self.num_layers = (self.log_size - 2) * 2 + 1

	self.convs = nn.ModuleList()
	self.upsamples = nn.ModuleList()
	self.to_rgbs = nn.ModuleList()
	self.noises = nn.Module()

	in_channel = self.channels[4]

	for layer_idx in range(self.num_layers):
	res = (layer_idx + 5) // 2
	shape = [1, 1, 2 res, 2 res]
	self.noises.register_buffer(f'noise_{layer_idx}', torch.randn(*shape))