ERNIE
Source code and dataset for "ERNIE: Enhanced Language Representation with Informative Entities"
Reqirements:
- Pytorch>=0.4.1
- Python3
- tqdm
- boto3
- requests
- apex (If you want to use fp16, you should make sure the commit is
79ad5a88e91434312b43b4a89d66226be5f2cc98.)
Prepare Pre-train Data
Run the following command to create training instances.
# Download Wikidump
wget https://dumps.wikimedia.org/enwiki/latest/enwiki-latest-pages-articles.xml.bz2
# Download anchor2id
wget -c https://cloud.tsinghua.edu.cn/f/1c956ed796cb4d788646/?dl=1 -O anchor2id.txt
# WikiExtractor
python3 pretrain_data/WikiExtractor.py enwiki-latest-pages-articles.xml.bz2 -o pretrain_data/output -l --min_text_length 100 --filter_disambig_pages -it abbr,b,big --processes 4
# Modify anchors with 4 processes
python3 pretrain_data/extract.py 4
# Preprocess with 4 processes
python3 pretrain_data/create_ids.py 4
# create instances
python3 pretrain_data/create_insts.py 4
# merge
python3 code/merge.py
If you want to get anchor2id by yourself, run the following code(this will take about half a day) after python3 pretrain_data/extract.py 4
# extract anchors
python3 pretrain_data/utils.py get_anchors
# query Mediawiki api using anchor link to get wikibase item id. For more details, see https://en.wikipedia.org/w/api.php?action=help.
python3 pretrain_data/create_anchors.py 256
# aggregate anchors
python3 pretrain_data/utils.py agg_anchors
Run the following command to pretrain:
python3 code/run_pretrain.py --do_train --data_dir pretrain_data/merge --bert_model ernie_base --output_dir pretrain_out/ --task_name pretrain --fp16 --max_seq_length 256
We use 8 NVIDIA-2080Ti to pre-train our model and there are 32 instances in each GPU. It takes nearly one day to finish the training (1 epoch is enough).
Pre-trained Model
Download pre-trained knowledge embedding from Google Drive/Tsinghua Cloud and extract it.
tar -xvzf kg_embed.tar.gz
Download pre-trained ERNIE from Google Drive/Tsinghua Cloud and extract it.
tar -xvzf ernie_base.tar.gz
Note that the extraction may be not completed in Windows.
Fine-tune
As most datasets except FewRel don't have entity annotations, we use TAGME to extract the entity mentions in the sentences and link them to their corresponding entitoes in KGs. We provide the annotated datasets Google Drive/Tsinghua Cloud.
tar -xvzf data.tar.gz
In the root directory of the project, run the following codes to fine-tune ERNIE on different datasets.
FewRel:
python3 code/run_fewrel.py --do_train --do_lower_case --data_dir data/fewrel/ --ernie_model ernie_base --max_seq_length 256 --train_batch_size 32 --learning_rate 2e-5 --num_train_epochs 10 --output_dir output_fewrel --fp16 --loss_scale 128
# evaluate
python3 code/eval_fewrel.py --do_eval --do_lower_case --data_dir data/fewrel/ --ernie_model ernie_base --max_seq_length 256 --train_batch_size 32 --learning_rate 2e-5 --num_train_epochs 10 --output_dir output_fewrel --fp16 --loss_scale 128
TACRED:
python3 code/run_tacred.py --do_train --do_lower_case --data_dir data/tacred --ernie_model ernie_base --max_seq_length 256 --train_batch_size 32 --learning_rate 2e-5 --num_train_epochs 4.0 --output_dir output_tacred --fp16 --loss_scale 128 --threshold 0.4
# evaluate
python3 code/eval_tacred.py --do_eval --do_lower_case --data_dir data/tacred --ernie_model ernie_base --max_seq_length 256 --train_batch_size 32 --learning_rate 2e-5 --num_train_epochs 4.0 --output_dir output_tacred --fp16 --loss_scale 128 --threshold 0.4
FIGER:
python3 code/run_typing.py --do_train --do_lower_case --data_dir data/FIGER --ernie_model ernie_base --max_seq_length 256 --train_batch_size 2048 --learning_rate 2e-5 --num_train_epochs 3.0 --output_dir output_figer --gradient_accumulation_steps 32 --threshold 0.3 --fp16 --loss_scale 128 --warmup_proportion 0.2
# evaluate
python3 code/eval_figer.py --do_eval --do_lower_case --data_dir data/FIGER --ernie_model ernie_base --max_seq_length 256 --train_batch_size 2048 --learning_rate 2e-5 --num_train_epochs 3.0 --output_dir output_figer --gradient_accumulation_steps 32 --threshold 0.3 --fp16 --loss_scale 128 --warmup_proportion 0.2
OpenEntity:
python3 code/run_typing.py --do_train --do_lower_case --data_dir data/OpenEntity --ernie_model ernie_base --max_seq_length 128 --train_batch_size 16 --learning_rate 2e-5 --num_train_epochs 10.0 --output_dir output_open --threshold 0.3 --fp16 --loss_scale 128
# evaluate
python3 code/eval_typing.py --do_eval --do_lower_case --data_dir data/OpenEntity --ernie_model ernie_base --max_seq_length 128 --train_batch_size 16 --learning_rate 2e-5 --num_train_epochs 10.0 --output_dir output_open --threshold 0.3 --fp16 --loss_scale 128
Some code is modified from the pytorch-pretrained-BERT. You can find the explanation of most parameters in pytorch-pretrained-BERT.
As the annotations given by TAGME have confidence score, we use --threshlod to set the lowest confidence score and choose the annotations whose scores are higher than --threshold. In this experiment, the value is usually 0.3 or 0.4.
The script for the evaluation of relation classification just gives the accuracy score. For the macro/micro metrics, you should use code/score.py which is from tacred repo.
python3 code/score.py gold_file pred_file
You can find gold_file and pred_file on each checkpoint in the output folder (--output_dir).
New Tasks:
If you want to use ERNIE in new tasks, you should follow these steps:
- Use an entity-linking tool like TAGME to extract the entities in the text
- Look for the Wikidata ID of the extracted entities
- Take the text and entities sequence as input data
Here is a quick-start example (code/example.py) using ERNIE for Masked Language Model. We show how to annotate the given sentence with TAGME and build the input data for ERNIE. Note that it will take some time (around 5 mins) to load the model.
# If you haven't installed tagme
pip install tagme
# Run example
python3 code/example.py
Cite
If you use the code, please cite this paper:
@inproceedings{zhang2019ernie,
title={{ERNIE}: Enhanced Language Representation with Informative Entities},
author={Zhang, Zhengyan and Han, Xu and Liu, Zhiyuan and Jiang, Xin and Sun, Maosong and Liu, Qun},
booktitle={Proceedings of ACL 2019},
year={2019}
}
110 Jan 07, 2023
5 Oct 29, 2022
66 Dec 31, 2022
47 Sep 10, 2022
4 Nov 24, 2021
4 Jul 08, 2022
9 Oct 31, 2022
66 Dec 14, 2022
138 Dec 27, 2022
103 Dec 10, 2022
317 Dec 23, 2022
29 Nov 30, 2022
14 Dec 15, 2021
409 Jan 06, 2023
21 Sep 25, 2022
101 Dec 26, 2022
12 Nov 05, 2022
1 Dec 17, 2021