Comments (4)
(Sorry for the delayed reply, been busy lately.)
Hey there! The results generated from AlexNet unfortunately don't look very nice. If you'd like to stick with AlexNet, I'd recommend 1) fine-tuning a new AlexNet model with all of the LRN layers nixed, and 2) decreasing the number of fully connected layers at the very end. You're free to decrease the stride of the layers, but do note that this will greatly increase your runtime, and you'll have to train the model entirely from scratch.
from style-transfer.
Thanks for your reply.
I have seen the result you generated from AlexNet in you blog. And I know the result should not be as good as VGG or Inception. So you get your result by removing the LRN layers? I just got terrible result which can't recognize the content image :(
from style-transfer.
Actually the results you saw were generated without removing LRN layers. If your results are unrecognizable, I'd recommend lowering the style-to-content ratio (maybe to 1e2
or 1e3
as a starting point).
If you paste an image here, I'd be happy to help you debug the problem.
from style-transfer.
Thank you so much.
I just got similar results as your blog using your code.
Appreciate your kindly help!
from style-transfer.
Related Issues (20)
- GPU Load 5% avg HOT 3
- Some about implementation HOT 2
- Issue about Loss function HOT 1
- Why do backpropagation for original VGG net? HOT 1
- i try to use it, but not working
- The program stops at "successfully loaded images" HOT 7
- training script
- where the parameter ‘n_iter’ affects in the minimize() function
- style doesn't tranformed successfully but only content HOT 1
- Add description for OpenCL, works well as replacement for CUDA HOT 2
- Check failed ! HOT 1
- why net to flatten img0? HOT 1
- CPU only? HOT 4
- Support for multiple style images HOT 1
- Model.ckpt
- On Windows, minimisation fails
- style.py: AttributeError: 'module' object has no attribute 'set_device' HOT 3
- Some way to set the standard output to lossless .png? HOT 1
- No module named 'caffe' HOT 2
- Segmentation fault
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 style-transfer.