Hello, I’m running the code you provided and noticed that it uses the Spider 1.0 dev split by default. I would like to reproduce the 86.2% accuracy on the Spider leaderboard using GPT-4 during testing.
Should the parameters I use be exactly the same as those in the run_dail_sql.sh script?
Are there any additional preprocessing steps or parameter adjustments needed?
Thank you for your great work and for open-sourcing the code. I really appreciate any guidance or corrections you can provide regarding my setup!
Hello, I’m running the code you provided and noticed that it uses the Spider 1.0 dev split by default. I would like to reproduce the 86.2% accuracy on the Spider leaderboard using GPT-4 during testing.
Should the parameters I use be exactly the same as those in the run_dail_sql.sh script?
Are there any additional preprocessing steps or parameter adjustments needed?
Thank you for your great work and for open-sourcing the code. I really appreciate any guidance or corrections you can provide regarding my setup!