awslabs / gap-text2sql

GAP-text2SQL: Learning Contextual Representations for Semantic Parsing with Generation-Augmented Pre-Training
https://arxiv.org/abs/2012.10309
Apache License 2.0
102 stars 25 forks source link

Query formed for total irrelevant input string #16

Open Mohdwajtech opened 3 years ago

Mohdwajtech commented 3 years ago

Hi, I was trying this on a dummy database attached with couple of tables. I observed that for a totally irrelevant input text it still outputs a query. Can we have a confidence score too along with the query in the response so that we can look into the confidence value and decide how relevant it is. Attached a screenshot of the issue too sqlprblm

alan-ai-learner commented 3 years ago

How you managed to run this notebook @Mohdwajtech, i unable to run it every time i face some errors.

Pls Help!

Impavidity commented 3 years ago

There is a score field in the inference output such as ie_dirs/bart_run_1_true_1-step41000.infer. This score might be helpful for helping you to identify low-confidence prediction.

Mohdwajtech commented 3 years ago

@Impavidity Thanks for your response. It helps. One more question I had. In case when I am running this on my own database, it is giving me error while I am adding more than 16 tables. The error I have shared below in the picture. Could you help me with this?

text2sql_error