[KongchangAI]
Unverified50% confidenceFactExact time

DeepSeek-V2提出MLA(Multi-head Latent Attention),通过低秩压缩将KV Cache压缩至MHA的5%左右

1
Sources
50%
Confidence
Long-term
Relevance
7/15/2026
First Seen

Sources

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/512131
API
curl https://kongchang.com/api/v1/knowledge/claims/512131
MCP
get_claim(id=512131)