Skip to content

Latest commit

 

History

History

README.md

This directory contains the code for the Parallel Cross Encoder method.

The process of training and testing the model is as follows:

First preprocess the dataset:

bash scripts/preprocessing_finsql.sh

Then train the Cross-Encoder model:

bash scripts/train_text2sql_schema_item_classifier_finsql.sh

At last, use the Cross-Encoder model to predict the dev set:

bash scripts/generate_text2sql_dataset_finsql.sh