Are raw Dolmino Mix 1124 compositions (50B, 100B and 300B) publicly available?

[baon6052]

❓ The question

Hi!

Is the raw data for the 50B, 100B and 300B compositions available for Dolmino Mix 1124?

I can see the tokenized .npy files are available to download here, but these have already been tokenized.

Thanks!

[ZeguanXiao]

I have the same need.

[baileykuehl]

Hi, thanks for the inquiry! This is something we did consider, but unfortunately, we don't have the raw data sectioned out into these specific compositions available currently. We have the compositions of the tokenized files, for example the 50B mix. But I've noted down this interest for future releases!