jeffwen / road_building_extraction Goto Github PK
View Code? Open in Web Editor NEWUsing Pytorch implementation of U-Net architecture for road and building extraction
License: MIT License
Using Pytorch implementation of U-Net architecture for road and building extraction
License: MIT License
Hi,
Thank you for uploading such great work.
I was wondering if this model can be used to add roads to OSM. As the training data is obtained from OSM itself, do you think we can use this model to enrich OSM ?
Thanks for posting your code. It's helping me wrap my head around neural networks and pytorch.
My task is satellite image classification. I have a 10 band multi spectral image that I've padded and gridded into 128x128 images. The last band is the class (0-6). I modified your data loader (really nice job by the way) so I have a 9x128x128 tensor for inputs and a 128x128 tensor for labels. I modified the model to take in 9 bands
# encoding
self.conv1 = encoding_block(9, 32)
My training and validation tensors appear to be correct when training the model. I checked this by printing out a small section from both sat and map in the train method.
My small test dataset has 63 images for training, 7 for validation, and 16 for testing. So far I'm able to set my batch size to 32 (8gb GTX1070) which means I only have two iterations per epoch.
I'm happy that it's running but every epoch produces the same loss and accuracy.
Epoch 71/99
----------
training: 100%|██████████████████████████████████████████████| 2/2 [00:02<00:00, 1.74s/it]
Training Loss: 1046.7444 Acc: 0.0000
validation: 100%|████████████████████████████████████████████| 3/3 [00:01<00:00, 1.57it/s]
Validation Loss: 693.5661 Acc: 0.0000
Current elapsed time 5m 33s
Any advice would be greatly appreciated.
Thank!
What is the format of the data read in the CSV file?
I have downloaded the datasets,but I don't find the datasets of "roads and crops",so where can Idownload the "roads and crops" datasets ?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.