slhaf/Partner
1
0
Fork 0
mirror of https://github.com/slhaf/Partner.git synced 2026-05-12 08:43:02 +08:00
Code Issues Packages Projects Releases Wiki Activity

推进 ActionExtractor 语义缓存机制: 两种嵌入模型的连接方式测试完毕，在高性能主机上，可以通过ollama调用mxbai-embed-large这类模型，但放到4核8G香橙派3B就会出现推理时长过长，哪怕换成ONNX RUNTIME JAVA 也难以避免，但如果更换成 nomic-embed-text + ONNX RUNTIME JAVA ，仍能够拿到70左右ms的推理时长，远低于提取模型以及向量模型API的调用时长。预期可提供两种语义缓存所用的嵌入模型接入方式: 通过 http 调用 本地ollama接口; 指定 ONNX 格式的嵌入模型直接调用。

Browse Source
This commit is contained in:
slhafzjw
2025-10-16 23:04:41 +08:00
parent e78048f66d
commit 7094a8a68b
7 changed files with 230 additions and 38 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
.idea/vcs.xml generated
View File

@@ -2,6 +2,5 @@
<project version="4">
<component name="VcsDirectoryMappings">
<mapping directory="$PROJECT_DIR$" vcs="Git" />
<mapping directory="$USER_HOME$/Projects/IdeaProjects/Projects/Partner" vcs="Git" />
</component>
</project>