python run_entropy.py -i INPUT_TEXT_FILE -o OUTPUT_NLL_FILE --model_path MODEL_NAME
NLL
is for negative log likelihood.
python run_fft.py -i INPUT_NLL_FILE -o OUTPUT_FFT_FILE -p zscore --value norm
You can set -p
to zscore
, logzs
, or other values for different normalization methods
- Use
circular.ipynb
to conduct circularization operation on likelihood scores. - Use
classifier_circlemean.ipynb
for supervised classification. - Use
classifier_pairwise.ipynb
for heuristic-based classification.