Hugging face already reproduced deepseek R1 (called Open R1) and open sourced the entire pipeline

The model weights and research paper are, which is the accepted terminology nowadays.

It would be nice to have the training corpus and RLHF too.

I’ve seen a lot of communities shift instances and change names. At least that seems to work here, so I wouldn’t feel too bad about disrespecting the unmoderated ones.

Obviously the active ones I would stick with

I was at a subaru dealership and they had a mid 2000s Subaru next to a new one

The old one was awesome, big windows, good storage, small footprint. It was way smaller than the modern one, which is a full on SUV.

I hate this.

I just bought a new battery for my 12 mini, I’m not going to upgrade until small phones are back.

But also this phone still works perfectly, it’s like the day I bought it after the new battery.

I left when they killed Apollo.

I never liked the native app, and I found lemmings to be good folk.

I think now is the most important time to ditch American social media.

I’ve started dining decaf for this reason

Never thought I would be a decaf drinker

I cannot watch that scene

Yes, there are LLMs for that, you literally just have to Google “llm parse PDF”.

You could also use tesseract or any number of other solutions which probably work as well…

But an inexperienced kid is gonna act like an inexperienced kid

Did you know python for loops have an else clause?

Did you also know it runs after the loop iterator is done, and not when there was never anything to iterate over at all?

https://docs.python.org/3/reference/compound_stmts.html#for