Unverified50% confidenceFactExact time
all-MiniLM-L6-v2 is an open-source sentence embedding model of about 80MB, often chosen by local applications as the default option, and Bees' local memory feature relies on it to encode text into vectors for storage and retrieval.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
Unverified主流嵌入模型如all-MiniLM-L6-v2、BGE系列、E5系列通常基于BERT架构，将文本编码为384至1536维的稠密浮点向量65% similarUnverifiedThe open-source project llama.cpp makes it possible to run quantized language models on devices with only a few GB of memory, such as a Raspberry Pi.64% similarUnverifiedMiniCPM5-2B is a small language model with 2B parameters.62% similarUnverified微调VLM能够复用大模型已有的视觉理解与语言生成能力，在相对有限的标注数据下实现对特定文献类型的适配62% similarUnverifiedvLLM is designed specifically for large language models and improves LLM throughput and GPU memory utilization through technologies such as PagedAttention.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1047152API
curl https://kongchang.com/api/v1/knowledge/claims/1047152MCP
get_claim(id=1047152)