Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/huybery/r2sql
:hot_pepper: R²SQL: "Dynamic Hybrid Relation Network for Cross-Domain Context-Dependent Semantic Parsing." (AAAI 2021)
https://github.com/huybery/r2sql
semantic-parsing text2sql
Last synced: about 1 month ago
JSON representation
:hot_pepper: R²SQL: "Dynamic Hybrid Relation Network for Cross-Domain Context-Dependent Semantic Parsing." (AAAI 2021)
- Host: GitHub
- URL: https://github.com/huybery/r2sql
- Owner: huybery
- Created: 2021-01-05T08:55:38.000Z (almost 4 years ago)
- Default Branch: master
- Last Pushed: 2021-12-20T08:53:41.000Z (about 3 years ago)
- Last Synced: 2024-08-03T09:07:27.999Z (5 months ago)
- Topics: semantic-parsing, text2sql
- Language: Python
- Homepage: https://arxiv.org/abs/2101.01686
- Size: 874 KB
- Stars: 61
- Watchers: 5
- Forks: 11
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- StarryDivineSky - huybery/r2sql - Domain Context-Dependent Semantic Parsing 跨域上下文相关语义分析的动态混合关系网络 应用于：多轮text-to-SQL 任务（通过多轮对话的方式生成最终的查询语句， Text-to-SQL 任务：给定一个自然语言查询和数据库的作为输入，产生一个SQL语句作为输出。） (其他_NLP自然语言处理 / 其他_文本生成、文本对话)
README
# R²SQL
The PyTorch implementation of paper [Dynamic Hybrid Relation Network for Cross-Domain Context-Dependent Semantic Parsing.](https://arxiv.org/pdf/2101.01686) (AAAI 2021)
## Requirements
The model is tested in python 3.6 with following requirements:
```
torch==1.0.0
transformers==2.10.0
sqlparse
pymysql
progressbar
nltk
numpy
six
spacy
```
All experiments on SParC and CoSQL datasets were run on NVIDIA V100 GPU with 32GB GPU memory.
* Tips: The 16GB GPU memory may appear out-of-memory error.
## Setup
The SParC and CoSQL experiments in two different folders, you need to download different datasets from [[SParC](https://yale-lily.github.io/spider) | [CoSQL](https://yale-lily.github.io/cosql)] to the `{sparc|cosql}/data` folder separately.
Another related data file could be download from [EditSQL](https://github.com/ryanzhumich/editsql/tree/master/data).
Then, download the database sqlite files from [[here](https://drive.google.com/file/d/1a828mkHcgyQCBgVla0jGxKJ58aV8RsYK/view?usp=sharing)] as `data/database`.
Download Pretrained BERT model from [[here](https://drive.google.com/file/d/1f_LEWVgrtZLRuoiExJa5fNzTS8-WcAX9/view?usp=sharing)] as `model/bert/data/annotated_wikisql_and_PyTorch_bert_param/pytorch_model_uncased_L-12_H-768_A-12.bin`.
Download Glove embeddings file (`glove.840B.300d.txt`) and change the `GLOVE_PATH` for your own path in all scripts.
Download Reranker models from [[SParC reranker](https://drive.google.com/file/d/1cA106xgSx6KeonOxD2sZ06Eolptxt_OG/view?usp=sharing) | [CoSQL reranker](https://drive.google.com/file/d/1UURYw15T6zORcYRTvP51MYkzaxNmvRIU/view?usp=sharing)] as `submit_models/reranker_roberta.pt`, besides the roberta-base model could download from [here](https://drive.google.com/file/d/1LkTe-Z0AFg2dAAWgUKuCLEhSmtW-CWXh/view?usp=sharing) for `./[sparc|cosql]/local_param/`.
## Usage
Train the model from scratch.
```bash
./sparc_train.sh
```
Test the model for the concrete checkpoint:
```bash
./sparc_test.sh
```
then the dev prediction file will be appeared in `results` folder, named like `save_%d_predictions.json`.
Get the evaluation result from the prediction file:
```bash
./sparc_evaluate.sh
```
the final result will be appeared in `results` folder, named `*.eval`.
Similarly, the CoSQL experiments could be reproduced in same way.
---
You could download our trained checkpoint and results in here:
* SParC: [[log](https://drive.google.com/file/d/19ySQ_4x3R-T0cML2uJQBaYI2EyTlPr1G/view?usp=sharing) | [results](https://drive.google.com/file/d/12-kTEnNJKKblPDx5UIz5W0lVvf_sWpyS/view?usp=sharing)]
* CoSQL: [[log](https://drive.google.com/file/d/1QaxM8AUu3cQUXIZvCgoqW115tZCcEppl/view?usp=sharing) | [results](https://drive.google.com/file/d/1fCTRagV46gvEKU5XPje0Um69rMkEAztU/view?usp=sharing)]
### Reranker
If your want train your own reranker model, you could download the training file from here:
* SParC: [[reranker training data](https://drive.google.com/file/d/1XEiYUmDsVGouCO6NZS1yyMkUDxvWgCZ9/view?usp=sharing)]
* CoSQL: [[reranker training data](https://drive.google.com/file/d/1mzjywnMiABOTHYC9BWOoUOn4HnokcX8i/view?usp=sharing)]
Then you could train, test and predict it:
train:
```bash
python -m reranker.main --train --batch_size 64 --epoches 50
```
test:
```bash
python -m reranker.main --test --batch_size 64
```
predict:
```bash
python -m reranker.predict
```
## Improvements
We have improved the origin version (descripted in paper) and got more performance improvements :partying_face:!
Compare with the origin version, we have made the following improvements：
* add the self-ensemble strategy for prediction, which use different epoch checkpoint to get final result. In order to easily perform this strategy, we remove the task-related representation in Reranker module.
* remove the decay function in DCRI, we find that DCRI is unstable with decay function, so we let DCRI degenerate into vanilla cross attention.
* replace the BERT-based with RoBERTa-based model for Reranker module.
The final performance comparison on dev as follows:
SParC
CoSQL
QM
IM
QM
IM
EditSQL
47.2
29.5
39.9
12.3
R²SQL v1 (origin paper)
54.1
35.2
45.7
19.5
R²SQL v2 (this repo)
54.0
35.2
46.3
19.5
R²SQL v2 + ensemble
55.1
36.8
47.3
20.9
## Citation
Please star this repo and cite paper if you want to use it in your work.
## Acknowledgments
This implementation is based on ["Editing-Based SQL Query Generation for Cross-Domain Context-Dependent Questions"](https://github.com/ryanzhumich/editsql) EMNLP 2019.