Repository for Adaptive Fusion Techniques for Multimodal Data (Python package to be released soon)
Make sure that you are in the root directory before running the following commands:
- Modify the
src/models/config.py
file appropriately before running any commands - If it's your first run (starting with nothing but the raw files), run
python src/process_raw.py
. This will create processed dataset files insidedata/processed
- Once you have the processed files, run
python src/train.py
to train the network. It will take care of:- Preprocessing the text
- Creating/Loading vocabulary
- Generating/Loading embeddings
- Training the model