firashm/gl-gin

GL-GIN: Fast and Accurate Non-Autoregressive Model for Joint Multiple Intent Detection and Slot Filling

This repository contains the official PyTorch implementation of the paper: Libo Qin, Fuxuan Wei, Tianbao Xie, Xiao Xu, Wanxiang Che, Ting Liu.

If you use any source codes or the datasets included in this toolkit in your work, please cite the following paper. The bibtex are listed below:

@misc{qin2021glgin,
      title={GL-GIN: Fast and Accurate Non-Autoregressive Model for Joint Multiple Intent Detection and Slot Filling}, 
      author={Libo Qin and Fuxuan Wei and Tianbao Xie and Xiao Xu and Wanxiang Che and Ting Liu},
      year={2021},
      eprint={2106.01925},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

In the following, we will guide you how to use this repository step by step.

Architecture

Results

Preparation

Our code is based on PyTorch 1.2 Required python packages:

• numpy==1.19.1
• tqdm==4.50.0
• pytorch==1.2.0
• python==3.6.12
• cudatoolkit==9.2
• fitlog==0.7.1
• ordered-set==4.0.2

We highly suggest you using Anaconda to manage your python environment.

How to run it

The script train.py acts as a main function to the project, you can run the experiments by the following commands.

# MixATIS_clean dataset (ON GeForce RTX2080TI)
python train.py -g -bs=16 -dd=./data/MixATIS_clean  -sd=./save/MixATIS_clean -nh=4 -wed=128 -ied=128 -ehd=256 -sdhd=128 -dghd=64 -nldg=2 -sgw=2 -ne=200

# MixSNIPS_clean dataset (ON TITAN Xp)
python train.py -g -bs=16 -dd=./data/MixSNIPS_clean  -sd=./save/MixSNIPS_clean -nh=8 -wed=64 -ied=128 -ehd=256 -sdhd=128 -dghd=128 -nldg=2 -sgw=1 -ne=100

You can directly load the best models we saved:

# MixATIS_clean dataset
python train.py -g -ne=0 -dd=./data/MixATIS_clean -sd=./save/MixATIS_best

# MixSNIPS_clean dataset
python train.py -g -ne=0 -dd=./data/MixSNIPS_clean -sd=./save/MixSNIPS_best

If you have any question, please issue the project or email me or fuxuanwei and we will reply you soon.