search

AlibabaResearch / DAMO-ConvAI

DAMO-ConvAI: The official repository which contains the codebase for Alibaba DAMO Conversational AI.
MIT License
1.08k stars 176 forks source link

Missing text2sql pairs in train folder #147

Open HelenGuohx opened 2 months ago

HelenGuohx commented 2 months ago

Hi, I want to express my appreciation for your outstanding work on text2sql research.

I recently downloaded the BIRD-bench dataset from https://bird-bench.github.io/ and noticed that the train folder seemed to be missing the text2sql pairs. I found the database descriptions and sqlite files, but not the actual text prompts and corresponding SQL queries.

However, I was glad to see that the dev.json file in the dev folder contains the text2sql pairs I was looking for.

Could you please clarify if the text2sql pairs are intentionally excluded from the train folder, or if there might be a missing file I should download?

superctj commented 4 weeks ago

Any update on this issue? I am also looking for SQL queries of the training set.

bird-bench commented 3 weeks ago

@HelenGuohx @superctj Thanks for interests in our work. Could you check whether you met connection errors or somehow? We re-downloaded again, and it seems train.json exists. For your convenience, I also attach this train.json here. Thanks. train.json

superctj commented 3 weeks ago

Thank you for your quick response! I appreciate it. btw, I didn't run into connection errors and somehow train.json is not in the decompressed directory.