pickxiguapi / Clean-Offline-RLHF

Offline RLHF codebase implementation for "Uni-RLHF: Universal Platform and Benchmark Suite for Reinforcement Learning with Diverse Human Feedback" (ICLR2024)
https://uni-rlhf.github.io/
MIT License
31 stars 2 forks source link

About the raw dataset of smarts #2

Closed TU2021 closed 4 months ago

TU2021 commented 4 months ago

Thank you for opening source for the good and solid work!

It seems that the raw dataset is collected by the author themselves, but I didn't find a pkl file for the raw dataset in the code. Could you provide the raw dataset pkl or the collection code?

Thanks again for your excellent work!

pickxiguapi commented 4 months ago

Sorry for late response. Due to my oversight, I missed uploading the smarts dataset. I will complete it by this Thursday (UTC+8).

pickxiguapi commented 4 months ago

Dataset Link: https://drive.google.com/file/d/1_KNH8EziubY2s3r6ySSSzLKnlXexInPp/view?usp=sharing. I will close the issue and feel free to reopen it if you have more questions.