A BERT-based Chinese NL2SQL Model

The Bert model is a large-scale corpus pre-training model based on Trans-former developed by Google.It has achieved considerable breakthroughs in al-most all sub-fields in the field of natural languageprocessing.Based on the Bert model,this paper designs a new model for the Chinese NL2SQL dataset for a more pratical task in the field of artificial intelligence:NL2SQL,which is the "natural language to SQL statement" task.A new model for the Chinese NL2SQL dataset was designed and acheieved an accuracy rate of more than 86%.