243006306
Results
3
issues of
243006306
I think it is necessary to adapt to future multi vector column scenarios in advance, for the following reasons: 1. More and more vector databases support multi vector columns, and...
enhancement
P2
P3
{"4":0.055145263671875,"6":0.01322174072265625,"16":0.012908935546875,"37":0.012176513671875,xxxxxxxxxx} 相比于BM25稀疏算法得出来的分数，BGE-M3模型稀疏出来的向量分数大多在0.01以内，存入milvus内通过稀疏检索，哪怕是使用同一稀疏向量的值去检索，得出的分值也在0.02分以内，这会让我排序阈值不好控制，请问是我缺少什么参数吗？通用RAG对于稀疏向量的阈值分数是多少呢？