How the dataset will be splitted

[riddiculous]

I'm interested in how the dataset will be splitted after the full version is released, and have a few questions.

1. Do you permit any training? Which part of the dataset can be used for training?
2. Do you have an extra test set without gold SQL on the leaderboard?

[lfy79001]

Hello,

Thank you for your interest in Spider 2.0. We will be releasing the paper and all the data in about 10 days. At that time, we will also launch a website to maintain the leaderboard. You can submit your results by email, which relies entirely on academic integrity as we do not have an anti-cheating mechanism in place.

Please note that Spider 2.0 does not have a train set or dev set, only a test set. We will provide the complete set of gold SQL queries and corresponding answers.