Welcome

landmark-detection

Implementation of landmark detection using keras.

You can find the project here

zld-scraper

Zalando scraper

Scrap images from zalando and train a classifier on them.

You can find the project here

Here are some of them:

CNN-layer-visualization

Web app for filters visualisation

An app to show the image of maximization of the filter’s activation.

You can find the project here

You just have to choose the layer, then put the filter number, it will automatically save it to static/img/layers_name/

If you put no filter number, it will show you all the filters of the layer.

Here are some examples :

All you have to do, is to replace the code of the function make_model() by your model architecture, in visualize.py, put your model in the folder, and change the model’s name in app.py, import also all the necessary dependencies for your model.

To run flask, change the path to app.py in run.sh, execute the command:

source run.sh

then:

flask run --port 5000 --host 0.0.0.0

Make sure to export the path to app.py before running the app

export FLASK_APP='root/your_path'