Comments (1)
Managed to bring variable (height,width) support in my case by changing the following lines in model.py for (1024,512). Model Wanted (1024,1024). I believe one could do a future PR and incorporate this permanently in the code.. Interesting feature to add.
For Input:
self.input = nn.Parameter(torch.randn(1, channel, size, size ))
to
self.input = nn.Parameter(torch.randn(1, channel, size, (size//2) ))
For noise Layers:
for layer_idx in range(self.num_layers):
res = (layer_idx + 5) // 2
#edited
shape = [1, 1, 2 ** res, 2 ** res]
to
shape = [1, 1, 2 ** res, (2 ** res)//2]
def make_noise(self):
device = self.input.input.device
#edited
noises = [torch.randn(1, 1, 2 ** 2, 2 ** 2, device=device)]
to
noises = [torch.randn(1, 1, 2 ** 2, (2 ** 2)//2, device=device)]
for i in range(3, self.log_size + 1):
for _ in range(2):
#edited
noises.append(torch.randn(1, 1, 2 ** i, 2 ** i, device=device))
to
noises.append(torch.randn(1, 1, 2 ** i, (2 ** i)//2, device=device))
But be forewarned if anyone tries this. This adds issues with the Wencoder which need to be retrained to match the dimensions.
from hyperstyle.
Related Issues (20)
- How to generate custom dataset ganspace pca.pt? HOT 1
- Could you please provide the classifier used for training interfacegan directions? HOT 3
- LSUN Bedrooms HOT 4
- About the 'smile/age/pose.pt' HOT 3
- about image inversion? HOT 1
- Pretrained Generator does not accept "weights_deltas" as a valid keyword HOT 2
- Training for different domain HOT 2
- About editing of AFHQW using StyleCLIP global-directions HOT 2
- Out-of-Domain Images HOT 1
- pretrained HyperStyle model gets wrong results. HOT 1
- About the Faces W-Encoder HOT 2
- Training on custom dataset. HOT 6
- Train new dataset based on pre-trained HyperStyle model HOT 1
- Pretrained W-Encoders pt file provided doesn't match the WEncoder class HOT 2
- WEncoder trained on e4e not working HOT 6
- 256px
- ModuleNotFoundError: No module named 'options'
- Very impressive work!!! Excuse me, the photo of the child, why are the results so bad? HOT 2
- How to maintain the cloth (neck) ?
- consult
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from hyperstyle.