https://github.com/svishnu88/TGS-SaltNet
Kaggle | 21st place solution for TGS Salt Identification Challenge
https://github.com/svishnu88/TGS-SaltNet
computer-vision deep-learning fastai kaggle-competition pytorch
Last synced: 3 months ago
JSON representation
Kaggle | 21st place solution for TGS Salt Identification Challenge
- Host: GitHub
- URL: https://github.com/svishnu88/TGS-SaltNet
- Owner: svishnu88
- Created: 2018-10-20T13:52:50.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2019-03-11T06:09:22.000Z (almost 6 years ago)
- Last Synced: 2024-08-08T23:22:46.114Z (7 months ago)
- Topics: computer-vision, deep-learning, fastai, kaggle-competition, pytorch
- Language: Python
- Homepage:
- Size: 27.3 KB
- Stars: 83
- Watchers: 7
- Forks: 24
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# TGS-SaltNet
Kaggle | 21st place solution for TGS Salt Identification Challenge
## General
I recently participated in a Kaggle competition [TGS Salt Identification Challenge](https://www.kaggle.com/c/tgs-salt-identification-challenge)
and reached the 21st place. This repository contains the final code which resulted in the best model. The code demonstrates usage of different important techniques using [fast.ai](http://www.fast.ai/) and [PyTorch](https://pytorch.org/).
1. Use ResNet model as an encoder for UNet.
2. Add intermediate layers like [BAM](http://bmvc2018.org/contents/papers/0092.pdf),[Squeeze & Excitation](https://arxiv.org/abs/1803.02579) blocks in a ResNet34 model which can be easily replicated for other network architectures.
3. Show how to add [Deep supervision](https://www.kaggle.com/c/tgs-salt-identification-challenge/discussion/65933) to the network, and calculate loss and combine loss at different scale.
## Main software used
1. fastai - 0.7
2. pytorch - 0.4
3. python - 3.6
## Hardware required
The code was tested with TitanX GPU/1080ti.
## Thanks
A special thanks to Heng for his generous contributions to different ideas in the competition, for a long list of amazing Kaglle community members, Jeremy and Fast.ai community for the amazing and flexible fastai framework.