ValueError：查询向量维度（384）与索引维度（768）不匹配
ValueError: Query vector dimension (384) does not match index dimension (768)
ID: llm/embedding-dimension-mismatch-in-vector-search
版本兼容性
| 版本 | 状态 | 引入 | 弃用 | 备注 |
|---|---|---|---|---|
| openai==1.10.0 | active | — | — | — |
| pinecone-client==3.0.0 | active | — | — | — |
| chromadb==0.4.22 | active | — | — | — |
| text-embedding-ada-002 | active | — | — | — |
| all-MiniLM-L6-v2 | active | — | — | — |
| sentence-transformers==2.2.2 | active | — | — | — |
根因分析
用于查询编码的嵌入模型与用于构建向量索引的模型输出维度不同，通常是由于模型切换（例如，从text-embedding-ada-002切换到较小的模型）。
English
Embedding model used for query encoding has a different output dimension than the model used to build the vector index, often due to switching between models (e.g., from text-embedding-ada-002 to a smaller model).
官方文档
https://platform.openai.com/docs/guides/embeddings解决方案
-
使用与查询相同的嵌入模型重新嵌入所有文档。例如，如果使用text-embedding-ada-002（1536维度），确保索引和查询都使用该模型。
-
使用维度无关的向量数据库（例如，带有自动检测功能的ChromaDB），通过存储元数据并在查询时过滤来处理不同维度。
无效尝试
常见但无效的做法:
-
90% 失败
Resizing the query vector by padding with zeros or truncating corrupts the embedding and leads to poor search results.
-
70% 失败
Re-indexing all documents with the new model is correct but often skipped due to time constraints; partial re-indexing causes inconsistency.
-
95% 失败
Assuming the vector database automatically handles dimension conversion leads to silent failures or errors.