This repository contains the source code to reproduce the experiments conducted in
Setup environment and install necessary packages / models:
To reproduce the main result from the paper simply run:
python evaluate.py
This will load the pre-computed predictions of several tools along with the gold labels.
Results are computed on the following corpora:
We obtiained the prediction and stored them in the pubtator format
in the annotations
folder for the following tools:
See the paper for more details.
You can obtain the predictions from our HunFlair2 tool with the following command:
python predict_hunflair2.py \
--input ./annotations/raw/bioid_text.txt \
--output ./annotations/hunflair2/bioid_text.txt \
--entity_types species
This will load a file in the PubTator formatand store the predictions into the --output
file