Bo Alpha

Darkdust's activity

• All
• Models
• Datasets
• Spaces
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

New activity in deepseek-ai/DeepSeek-R1 8 days ago

deepseek-r1 源码中采用 MLA 架构的 KV Cache 压缩存储策略的实现似乎与文中说的不一致，这是为什么？代码中似乎没实现这个大优化

2

#108 opened 10 days ago by

Darkdust

New activity in deepseek-ai/DeepSeek-R1 10 days ago

deepseek-r1 源码中采用 MLA 架构的 KV Cache 压缩存储策略的实现似乎与文中说的不一致，这是为什么？代码中似乎没实现这个大优化

2

#108 opened 10 days ago by

Darkdust