#trillion-parameter models

15 stories · 0 in 7 days · 1 in 30 days
Stories, products and papers tagged "trillion-parameter models", newest first.
Oct 29
Aug 12
Aug 11
Jul 31
Jul 21
Jul 20
Jul 14
Jul 10
Jul 8
Jun 10
Jun 9
Jun 8
ProductsUS vs ChinaOfficialSelected72°09:52
Chinese content昇腾910C完成1.6万亿参数模型全参数后训练,国产AI算力跨越关键门槛

国产算力终于跑通万亿参数模型全流程,做AI训练和模型研发的团队值得关注——这直接关系到未来能否摆脱对英伟达的依赖,建议点开了解技术细节。

Primary sourcePandaily@[email protected] (Pandaily)Source
May 28
ModelsUS vs ChinaOfficialSelected76°15:39
Chinese contentOrbit 开源框架:单节点训练万亿参数模型,DeepSeek-V4 可用

Orbit 让万亿参数模型的后训练不再依赖大规模集群,做 RL 微调或大模型优化的团队可以直接在单节点上跑 DeepSeek-V4,建议试试这个开源方案。

Primary sourcepandaily@[email protected] (Pandaily)Source