Thanks for open-sourcing such a great dataset! I am curious if you have done any pretraining from scratch using this dataset and what will be the recommended recipe. Thank you!
· Sign up or log in to comment
