Reproduce GPT-2 (124M) in llm.c in 90 minutes for $20

Re:Reproduce GPT-2 (124M) in llm.c in 90 minutes for $20 by ravetcofx on Tuesday May 28, 2024 @02:27PM (#40503867)

Holy crap, Does huggingface charge for bandwidth if you're downloading 45 terabytes??

[ Parent | on HN ]

Re:Reproduce GPT-2 (124M) in llm.c in 90 minutes for $20 by drexlspivey on Tuesday May 28, 2024 @02:47PM (#40504075)

I believe they are hosting it on Cloudflare who doesn’t charge for egress

[ Parent | on HN ]

Re:Reproduce GPT-2 (124M) in llm.c in 90 minutes for $20 by fragmede on Tuesday May 28, 2024 @03:02PM (#40504266)

More specifically, Cloudflare R2 doesn't charge for egress, and Cloudflare doesn't charge for egress to members in the Bandwidth Alliance which include Azure, Google Cloud, Oracle, Alibaba Cloud, and others, though critically not AWS.

They very much do charge egress fees elsewhere.

[ Parent | on HN ]

Re:Reproduce GPT-2 (124M) in llm.c in 90 minutes for $20 by andersa on Wednesday May 29, 2024 @12:21AM (#40508432)

Fun trivia: downloading 45TB costs about $60, according to Cloudflare.

[ Parent | on HN ]

Re:Reproduce GPT-2 (124M) in llm.c in 90 minutes for $20 by verticalscaler on Wednesday May 29, 2024 @07:17AM (#40510685)

That's what Cloudflare charges. It costs them around 6 cents.

[ Parent | on HN ]

Re:Reproduce GPT-2 (124M) in llm.c in 90 minutes for $20 by kazanz on Wednesday May 29, 2024 @12:56PM (#40514133)

Wish I could say I'm surprised you're getting downvotes. Carrier costs are some of the lowest costs for hosting providers. Yet that fact seems to elude a majority of the community here.

[ Parent | on HN ]

Re:Reproduce GPT-2 (124M) in llm.c in 90 minutes for $20 by andersa on Wednesday May 29, 2024 @05:36PM (#40517311)

That's what they said it costs on their blog, not that they charge that. https://blog.cloudflare.com/aws-egregious-egress

Where are you getting 6 cents from?

[ Parent | on HN ]