How has DeepSeek improved the Transformer architecture?

Article URL: https://epoch.ai/gradient-updates/how-has-deepseek-improved-the-transformer-architecture

Comments URL: https://news.ycombinator.com/item?id=42855170

Points: 121

# Comments: 30

https://epoch.ai/gradient-updates/how-has-deepseek-improved-the-transformer-architecture

созданный 1mo | 28 янв. 2025 г., 19:40:11

Войдите, чтобы добавить комментарий

Другие сообщения в этой группе

Chrome Returns 206 when the Server Returns 403

Chrome Returns 206 when the Server Returns 403

Article URL: https://aoli.al/blogs/chrome-bug/

Comments URL: https://news.ycombinator

3 мар. 2025 г., 21:50:11 | Hacker news

James Harrison, whose blood donations saved >2M babies, has died

James Harrison, whose blood donations saved >2M babies, has died

Article URL: https://www.npr.org/2025/03/03/nx-s1-5316163/james-harrison-blood-donor

Comments URL:

3 мар. 2025 г., 21:50:11 | Hacker news

One Logo, Three Companies

Article URL: https://estilofilos.blogspot.com/2016/03/one-logo-three-companies-i.html

Comments URL:

3 мар. 2025 г., 21:50:10 | Hacker news

Atlanta Fed predicts -2.8% GDP

Atlanta Fed predicts -2.8% GDP

Article URL: https://www.atlantafed.org/cqer/research/gdpnow

Comments URL:

3 мар. 2025 г., 21:50:10 | Hacker news

Tianjin robot incident raises alarm over public safety and robotics

Tianjin robot incident raises alarm over public safety and robotics

Article URL: https://www.businesskorea.co.kr/news/articleView.html?idxno=236533

Comments URL:

3 мар. 2025 г., 21:50:07 | Hacker news

Comparing Fuchsia components and Linux containers [video]

Comparing Fuchsia components and Linux containers [video]

Article URL: https://fosdem.org/2025/schedule/event/fosdem-2025-5381-comparing-

3 мар. 2025 г., 21:50:07 | Hacker news

Nomadic infrastructure design for AI workloads

Nomadic infrastructure design for AI workloads

Article URL: https://www.tigrisdata.com/blog/nomadic-compute/

Comments URL:

3 мар. 2025 г., 21:50:06 | Hacker news

Techie