MMNF-RepSeP
The Reproducible Self Publishing toolkit for the Master Thesis "Multi Modal Generative Learning utilizing Normalizing Flows".
Dependencies
- MMVAE_Hub
- pythontex (>=
0.16) - texlive-latex
- graphviz
Usage
Reproducing the original document
git clone git@github.com:Jimmy2027/MMNF_RepSeP.git
cd MMNF_RepSeP
./reproduce.sh
Launch model training and evaluation
The script to launch the model training was written to work on the GPU cluster leomed from ETH Zurich.
- Adapt the paths in the config under
prepare_thesis/conf.json python prepare_thesis/launch_model_training.py(this will launch the model training on leomed. This may take while.)- When all jobs have finished executing, do:
./produce.sh