How Meta trains large language models at scale

Re:How Meta trains large language models at scale by bluedino on Wednesday June 12, 2024 @10:57PM (#40665474)

We have almost 400 H100's sitting idle. I wonder how many other companies are buying millions of dollars worth of these chips with the hopes of them being used, but aren't being utilized?

[ Parent | on HN ]

Re:How Meta trains large language models at scale by jonathanlei on Thursday June 13, 2024 @12:07AM (#40665835)

Hello! If you're interested in monetizing those GPUs, I'd be happy to rent them (all 400!) and offer those to customers of the cloud I work at :)

jonathan [at] tensordock.com

[ Parent | on HN ]

Re:How Meta trains large language models at scale by jedberg on Thursday June 13, 2024 @12:32AM (#40665973)

If you want 512 H100s connected with infiniband: https://lambdalabs.com/service/gpu-cloud/1-click-clusters

[ Parent | on HN ]

Re:How Meta trains large language models at scale by radq on Wednesday June 12, 2024 @11:47PM (#40665750)

Have you considered sponsoring an open-source project? ;)

[ Parent | on HN ]

Re:How Meta trains large language models at scale by fragmede on Wednesday June 12, 2024 @11:16PM (#40665581)

the world would love to buy time on your idle H100's if you're selling.

[ Parent | on HN ]

Re:How Meta trains large language models at scale by irjustin on Wednesday June 12, 2024 @11:01PM (#40665494)

That's insane and incredible all at the same time.

[ Parent | on HN ]

Re:How Meta trains large language models at scale by gfosco on Wednesday June 12, 2024 @11:11PM (#40665553)

Looking to get rid of a few?....

[ Parent | on HN ]

Re:How Meta trains large language models at scale by newswasboring on Wednesday June 12, 2024 @11:07PM (#40665526)

Wow, that's a lot of money in inventory. What was the original thought process? Just fomo?

[ Parent | on HN ]

Re:How Meta trains large language models at scale by giancarlostoro on Wednesday June 12, 2024 @11:29PM (#40665647)

Probably could profit selling them second hand honestly.

[ Parent | on HN ]

Re:How Meta trains large language models at scale by Der_Einzige on Thursday June 13, 2024 @12:59AM (#40666115)

I know of many important projects that need GPUs right now and aren’t getting any. You could help motivate the ponydiffusion folks to actually try finetuning of SD3!

[ Parent | on HN ]

Re:How Meta trains large language models at scale by TeMPOraL on Thursday June 13, 2024 @02:26AM (#40666545)

So you're saying, H100s are the corporate equivalent of Raspberry Pis now? Bought to not miss out, then left to gather dust in a drawer?

[ Parent | on HN ]

Re:How Meta trains large language models at scale by mhsiu on Friday June 14, 2024 @12:09AM (#40677403)

If you'd like to contribute some much-appreciated compute to university researchers all across the US, please email me (mhsiu at ucsd.edu).

It should be possible to hook up your idle devices to Nautilus (https://nationalresearchplatform.org/), which is a cluster set up to support researchers at a bunch of universities. I can't guarantee anything since I'm not involved in the cluster management itself, but I can put you in contact with those who are if you're interested.

[ Parent | on HN ]