Content prediction
The software makes it possible to:
- train/improve the predictions models
- use trained model to perform predictions
There app comes with models pre-trained on the included data.
Usage
Training
The app comes pre-trained on the included data. Training is only needed if new data is added.
The data for training is provided in png format and is stored in the trainingset directory. The file name contains the content description, see examples.
There are 2 training methods available, they can be accessed by calling the following apps:
- python train.py - trains Resnet18
- python trainxgboost - trains XGBoostRegressor
Training produces model files that are later used in predictions.
Predictions
The prodiction is available through a web interface. The implementation uses Streamlit, therefore streamlit is required. To perform predictions run 'streamlit run predict.py' and upload images through the web interface. The predictions will be presented on the website, there is also an option to download them in Excel format.
Notes
SessionState used in the webinterface is taken from the following gist: https://gist.github.com/tvst/036da038ab3e999a64497f42de966a92