search

spring-projects / spring-ai

An Application Framework for AI Engineering
https://docs.spring.io/spring-ai/reference/1.0-SNAPSHOT/index.html
Apache License 2.0
2.52k stars 607 forks source link

neo4j Vector store does not allow more than 2048 embedding dimensions #816

Open tanjcf opened 1 month ago

tanjcf commented 1 month ago

Bug description neo4j vector store does not allow more than 2048 embedding dimensions. However I'm using llama3 8B as an embedding model so It produces vectors with a dimension of 4096.

Environment Spring AI 1.0-SNAPSHOT neo4j:5.20.0 (standalone docker)

Steps to reproduce Use neo4j as a vector store, setup a local instance of Milvus and add the starter to the pom.xml :

org.springframework.ai spring-ai-neo4j-store-spring-boot-starter ${spring-ai.version}

Run the application : defined in file [/Users/jctanking/workerpases/palette/palette-boot/target/classes/com/kd/palette/RagWorker.class]: Unsatisfied dependency expressed through constructor parameter 0: Error creating bean with name 'vectorStore' defined in class path resource [org/springframework/ai/autoconfigure/vectorstore/neo4j/Neo4jVectorStoreAutoConfiguration.class]: 'vector.dimensions' must be between 1 and 2048 inclusively Proposed solution image

tanjcf commented 1 month ago

image