Note: 5 Fold cross validation results are reported here. Mean and Std value as (mean ± std). H2O results are for 80:20 split, could not do kfold validation as it is compute and memory heavy.
- Create python environment and install dependencies from requirements.txt
- Put your
data
indata/
and starting running the notebook sequentially
- Make sure that you have
test
data indata/
and H2O model weights inmodels
- Check the path in the evaluate.py and make sure that they're correct
- Run evaluate.py