ELQ4GMTQA

Tailored ELQ for entity linking of GMT-KBQA

Easy Start

• Create an conda environment accroding to configuration file blink37.yml:

conda env create -f blink37.yml

• Activate environment:

conda activate blink37

• Download pretrained models: follow script from BLINK repository. By running the script, downloaded models will be saved in models/ by default.
• Start ELQ service, the default port is 5688:

CUDA_VISIBLE_DEVICES=3 python elq_service.py