FourierGPT

Estimate likelihood scores

python run_entropy.py -i INPUT_TEXT_FILE -o OUTPUT_NLL_FILE --model_path MODEL_NAME

NLL is for negative log likelihood.

Normalize likelihood and Fourier transform

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

Further steps

• 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.