Comments (3)
@Zephyr69
I'm sorry for the late reply.
That task may be difficult with the current implementation.
The implemented GAN's structure are not suitable for generating large images.
Please consider forking this repository and modifying the GAN's structure.
from f-anogan.
Hi, @Zephyr69.
Thanks for using this repository.
It is difficult to judge from your information alone. In your configuration, is image generation not working?
from f-anogan.
Hi, @Zephyr69. Thanks for using this repository.
It is difficult to judge from your information alone. In your configuration, is image generation not working?
Thanks for your reply! I'm looking to detect very small anomalies (~7x7) on a relatively big image (~1000x1000).
The image reconstruction is working with --latent_dim 100, but the image quality isn't good enough to perform the task. The G_loss quickly becomes all negative, idk if it's expected? Like this:
[Epoch 0/450] [Batch 0/2] [D loss: 9.225899] [G loss: 0.002858]
[Epoch 1/450] [Batch 0/2] [D loss: 8.284700] [G loss: 0.007262]
[Epoch 2/450] [Batch 0/2] [D loss: 5.510290] [G loss: -0.023316]
[Epoch 3/450] [Batch 0/2] [D loss: -4.197353] [G loss: -0.223050]
[Epoch 4/450] [Batch 0/2] [D loss: -32.129951] [G loss: -2.417758]
[Epoch 5/450] [Batch 0/2] [D loss: -88.610680] [G loss: 9.614916]
[Epoch 6/450] [Batch 0/2] [D loss: -126.860847] [G loss: 17.268265]
[Epoch 7/450] [Batch 0/2] [D loss: -106.419914] [G loss: -2.722338]
[Epoch 8/450] [Batch 0/2] [D loss: -115.030739] [G loss: -57.363503]
[Epoch 9/450] [Batch 0/2] [D loss: -0.776878] [G loss: -69.117149]
[Epoch 10/450] [Batch 0/2] [D loss: -19.475677] [G loss: -83.537827]
[Epoch 11/450] [Batch 0/2] [D loss: -303.673645] [G loss: 167.087326]
[Epoch 12/450] [Batch 0/2] [D loss: -146.033997] [G loss: -121.324600]
...
Increasing --latent_dim to 1000 will make image reconstruction fail; the generator won't produce any meaningful image.
from f-anogan.
Related Issues (12)
- Question about images_e encoder output HOT 2
- I want to train this f-anoGAN networks with my image datasets. HOT 8
- Has anyone succeeded in DAGM2007 dataset? HOT 1
- python setup.py install HOT 3
- How can I implement a validation dataset? HOT 5
- G_loss and D_loss are negative numbers when running train_wgangp.py using my own image dataset HOT 7
- Train AnoGAN on pcam dataset in h5 format HOT 3
- Need absolute value for image subtraction HOT 1
- wrong line in mvtec jupyter notebook HOT 3
- Graph values is off. HOT 4
- I want to train this f-anoGAN networks with my image datasets. HOT 1
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 f-anogan.