run the two demo.sh under src for getting the results with prototype and hierarichal label embeddings

uncompress data.tar.gz to ./Data to get the raw data for BBN

uncopmress intermediate.tar.gz to ./Intermediate to get the features extacted for BBN.

feaures used in this project is reused from the repository of https://github.com/shanzhenren/PLE. They recently released a new version called AFET, but I haven't tested the compatibility with our code.

#example: bash warp_zero_demo.sh 0.3 3 #threshold=0.3 max_depth=3 during inference.