Unverified50% confidenceSolutionExact time
The project ported a KV offloading approach similar to HiSparse to QSA (the sparse attention structure used by Qwen3-Next), keeping the full history in host memory while only retaining hot data in VRAM.
1
Sources
50%
Confidence
Long-term
Relevance
9/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe project's vector storage uses the Qdrant vector database.65% similarUnverified语义记忆的技术实现依赖向量数据库，本地部署方案包括ChromaDB、Qdrant和FAISS63% similarUnverifiedPersistent memory typically relies on vector databases for semantic retrieval, converting information into high-dimensional dense vectors via embedding models and storing them in systems such as Qdrant, Pinecone, Weaviate, and Chroma.63% similarVerifiedThe implementation of the semantic memory layer typically relies on vector databases (such as Chroma, Qdrant, and Pinecone) for vectorized storage and similarity retrieval, a technical approach known as RAG.62% similarPartially VerifiedHermes Agent uses vector databases (such as ChromaDB or Qdrant) to convert conversation history and knowledge fragments into high-dimensional vectors for storage and semantic retrieval62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/923183API
curl https://kongchang.com/api/v1/knowledge/claims/923183MCP
get_claim(id=923183)