Comments (5)
ok great, I will try that. Thank you
from trainner.
@mmmoussa96 Hello! Yours is a different situation, but it sounds potentially like an OpenCV BGR -> RGB conversion issue.
OpenCV by default reads images in BGR channel order, which are converted to RGB when converted to tensor for training.
Before line 158 here: https://github.com/victorca25/BasicSR/blob/master/codes/data/aligned_dataset.py, you can use the tmp_vis() debug function to visualize the images (uncomment line 9) and check if thatβs the problem.
The function should look something like:
tmp_vis(img_A, False)
Try that and let me know how they look like.
from trainner.
Hello! Maybe there's a missing conversion between RGB and OpenCV's BGR channel format. I would have to check where this could be happening, but can you try to do the upscaling using: https://github.com/victorca25/iNNfer instead? It should be easier to use and I've tested it a lot recently.
from trainner.
was this fixed? I'm having a similar issue (where it's green). I converted images to npy used their sub-images as per the guide here: https://github.com/victorca25/BasicSR/wiki/IO-speed
from trainner.
Yes, I used PIL to pre-process the images and I guess forgot to convert RGB -> BGR. Did that, trained the model and now I'm getting this blue image described. Will use iNNfer now. Thank you!
from trainner.
Related Issues (20)
- [Feature Request] Curriculum Training for Augmentations
- Pixel Unshuffle is broken HOT 2
- Video dataloader crashes at 1x scale
- Video learning rate too high
- Add lr_crop_size in config HOT 1
- Update requirements.txt with proper versioning for torch
- Feature request/bug fix: Perform scaling and other operations in linear light HOT 3
- `nearest_aligned` is not aligned HOT 2
- so many bugs in your sftgan implementation HOT 6
- How do i even use this HOT 2
- Is there any way to train video super resolution models using this? HOT 4
- lmdb has no valid image file HOT 3
- how to train for real ESRGAN HOT 1
- GPU usage at 0% during training HOT 7
- FutureWarning and UserWarning HOT 5
- I Have No Idea What I Am Doing to Cause This: HOT 5
- "cv2.error: OpenCV(4.7.0) D:\a\opencv-python\opencv-python\opencv\modules\imgproc\src\resize.cpp:4065: error: (-215:Assertion failed) inv_scale_x > 0 in function 'cv::resize'" HOT 1
- ETA is grossly over-estimated. HOT 2
- Pix2Pix 3->1 channel HOT 1
- AttributeError: module 'collections' has no attribute 'Iterable' HOT 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
π 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. πππ
-
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 β€οΈ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from trainner.