DeepSeek spends $294k to train R1 AI model – 亚太新闻周刊

Dark

Light

DeepSeek trained the R1 model using 512 Nvidia H800 chips, hardware designed for the Chinese market.

Tags

Related Posts

Nvidia spends $900m to hire AI startup Enfabrica CEO, staff

9 小时 ago

Tencent unveils Hunyuan3D 3.0 AI model with enhanced accuracy

3 天 ago

Alibaba, Baidu start using own chips to train AI models: sources

1 周 ago

《亚太新闻周刊》

发表回复取消回复

Latest from Blog

Both will partner on media content and host events highlighting innovation.

by 《亚太新闻周刊》

ABC pre-empts "Jimmy Kimmel Live!" over Kimmel's Charlie Kirk comments

by 《亚太新闻周刊》

Indonesian gov’t plans content filtering rules for Netflix, YouTube

by 《亚太新闻周刊》

HK to strengthen primary healthcare, deepen system reforms

by 《亚太新闻周刊》

The Puzzling Death of Susann Sills

by 《亚太新闻周刊》

Indian edtech firm Jaro Education to raise $50m in IPO

by 《亚太新闻周刊》

Human remains found in search for fugitive Travis Decker, accused of killing his 3 daughters in Washington state, officials say

by 《亚太新闻周刊》

US government’s Intel stake hits $14b after Nvidia deal

by 《亚太新闻周刊》

Dexie communicates in English, Mandarin, Malay, and several Chinese dialects.

by 《亚太新闻周刊》

It houses an epilepsy monitoring unit with 16 single-patient rooms.

by 《亚太新闻周刊》

Don't Miss

Nvidia spends $900m to hire AI startup Enfabrica CEO, staff

Nvidia had p

Tencent unveils Hunyuan3D 3.0 AI model with enhanced accuracy

Hunyuan 3D 3