View Code? Open in Web Editor
NEW
Colab files for running PhD work
colab-gists's Introduction
Within this repo are the 6 different colab notebooks used for Monument Monitor experimental AI work
- 1s First model made with tensorflow. Off the shelf model using semantic segmentation and ADE20k dataset
- 2s Second segmentation model - uses transfer learning
- 3s Third segmentation model - uses transfer learning with different configuration to model two
- 1c First classification model for water levels, using Alexnet
- 2c Second (better) classification model for water levels using pre-trained GoogleNet
- 3c This classification model, that looks creates classifier for all MM case study sites using images submitted up until 1st Feb 2021.
Model Number |
1s |
2s |
3s |
1c |
2c |
3c |
Architecture |
DeepLabv3+ |
ResNet101 |
ResNet101 |
AlexNet |
GoogLeNet |
ResNet18 |
Analysis |
type |
Segmentation |
Segmentation |
Segmentation |
Classification |
Classification |
Focus |
Area of water pooling |
Area of water pooling |
Area of water pooling |
Level of water pooling |
Level of water pooling |
Identification of site |
Validation and learning loss tracked |
False |
False |
True |
True |
True |
True |
![DOI](https://camo.githubusercontent.com/e04edcb29a71f24a784215da3c9b4e52966188645d1460c6f2d1ba5ac6fa5219/68747470733a2f2f7a656e6f646f2e6f72672f62616467652f39333134383732362e737667)
colab-gists's People
Contributors
Stargazers
Watchers