Shulou(Shulou.com)06/01 Report--

This article mainly introduces the improvement of Milvus v0.5.3 version, which is very detailed and has a certain reference value. Friends who are interested must finish reading it!

Major improvement

-the result set data transfer performance from Milvus server to the client has been at least doubled, mainly through the following updates to gRPC

Optimized messages

Changed the API interface that generated the code

Deleted compression

-Python SDK

Different arrays store ids and distances of search results separately to reduce the response time for the interface to read the result set

Added a new way to get a target vector in the result set: id = results.id_ array [I] [j], distance = results.distance_ array [I] [j]

A new array traversal method has been added, which greatly reduces the processing time in the case of large nq and topk.

> > for id_list, dis_list in zip (results.id_array, results.distance_array):

> > for id, dis in zip (id_list, dis_list):

> > print ("id= {}, distance= {}" .format (id, dis))

-Java SDK

When connecting to Milvus server, added setting options such as keepalive and idleTimeout.

Users can now get the ids and distances of search results through getResultIdsList and getResultDistancesList, respectively, for better performance. You can also get an object QueryResult that contains both ids and distances through getQueryResultsList.

-C++ SDK

Change it to a dynamic library

Added README file

-improved search performance of IVF_SQ8H

These are all the contents of this article entitled "what are the improvements in Milvus v0.5.3?" Thank you for reading! Hope to share the content to help you, more related knowledge, welcome to follow the industry information channel!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.