Closed HeegyuKim closed 1 month ago
Thank you very much for your interest in our work!
Regarding the issue you raised, I have successfully reproduced and compared the results. The discrepancy is due to a recent update to the Dev dataset by the BIRD team, which means it is not the same version as the one used in our paper.
I have uploaded the BIRD Dev dataset we used in our paper to Google Drive. Here is the link: Google Drive
I hope this resolves your issue.
Additionally, we have recently released the NL2SQL360 Package, which provides an easy-to-use, multi-angle solution. Feel free to use it! If you have any other questions, feel free to discuss them with us!
Hello! Thank you for sharing nice work!
I downloaded your published baseline results and evaluated them using the BIRD code. However, the BIRD benchmark result is different from the paper's.
I have been trying to solve this issue and I have noticed that my BIRD dev database filename is dev_20240627.zip. This different database version can be a reason for different reproduction results.
Could you share your dev dataset version?