Some questions about llama’s train data set

OpenBMB / ToolBench

[ICLR'24 spotlight] An open platform for training, serving, and evaluating large language model for tool learning.

https://openbmb.github.io/ToolBench/

Apache License 2.0

4.6k stars 397 forks source link

Some questions about llama’s train data set #257

Open zxthesky opened 3 months ago

zxthesky commented 3 months ago

Hi team,Sorry to bother you.

I am a student of Soochow University. My email address is [1805565641@qq.com]. I tried to use your method to preprocess the files in your 'answei' folder, but I couldn't get your toolllama_G123_dfs_train.json. For some queries in the file, the assistant is not trained every time. I want to know more. What are your screening criteria and how did you get the final toolllama_G123_dfs_train.json?

Thanks.

tttonytan commented 3 weeks ago

hi do you get this dataset? If you get, would you like to share it with me? 1656125392@qq.com

slatter666 commented 5 days ago

I guess the data has already preprocessed in Google Drive link https://drive.google.com/drive/folders/1yBUQ732mPu-KclJnuQELEhtKakdXFc3J, data.zip