We have a massive GPU cluster and developed our own infrastructure to manage the cluster and train massive models.
There's how it works:
- You upload the dataset with preconfigured format into HuggingFaсe [1].
Choose your LLM (e.g. LLaMa 70B, Mistral 7B)
- Place your submission into the queue
- Wait for it to get trained.
- Then you get your trained model there on HuggingFace.
Essentially, why would we want to do it?
We already have an experience with training big LLMs.
We could achieve near-perfect infrastructure performance for training.
Sometimes GPUs have just nothing to train.
Thus we thought it would be cool if we could utilize our GPU cluster 100%. And give back to Open Source community (already built an e2e distributed training framework [2]).
This is in an early stage, so you can expect some bugs.
Any thoughts, opinions, or ideas are quite welcome!
Software developer, whose passion is the creation of elegant interfaces with unmatched attention to details. I also understand the importance of creating highly readable and easily maintainable source code. I am constantly striving to learn new technologies and look to ways to better myself.
I live in Kazakhstan and this story is absolutely true in my opinion. Yerlan (Irlan is incorrect) just wanted to get "easy" money from foreign. I even think this was a first time when he'd wanted to take a bribe as he hadn't known how computer system works.
"Cops being happy to have some "criminal" around for getting drunk and they even pay?" - lol he was just a poor student whose visa expired, he wasn't a drug dealer. In Kazakhstan, most people are hospitable. And paying for guests is absolutely normal.
"The girl's story did not make much sense" - in our country sexual education is at a very low level. Abortions at young age is a big problem in poor small cities like where author have been.