--- id: siliconflow.md title: "SiliconFLow" summary: "This topic describes how to configure and use SiliconFLow embedding functions in Milvus." beta: Milvus 2.6.x --- # SiliconFLow This topic describes how to configure and use SiliconFLow embedding functions in Milvus. ## Choose an embedding model Milvus supports embedding models provided by SiliconFLow. Below are the currently available SiliconFLow embedding models for quick reference:
Model Name |
Dimensions |
Max Tokens |
Description |
|---|---|---|---|
BAAI/bge-large-zh-v1.5 |
1,024 |
512 |
A large Chinese text embedding model, which is part of the BGE (BAAI General Embedding) series. |
BAAI/bge-large-en-v1.5 |
1,024 |
512 |
A large English text embedding model, which is part of the BGE (BAAI General Embedding) series. |
netease-youdao/bce-embedding-base_v1 |
768 |
512 |
A bilingual and cross-lingual embedding model developed by NetEase Youdao. The model exhibits excellent performance in Chinese and English semantic representation and retrieval tasks, especially excelling in cross-lingual scenarios. |
BAAI/bge-m3 |
1,024 |
8,192 |
A multifunctional, multilingual, multi-granularity text embedding model. It supports three common retrieval functions: dense retrieval, multi-vector retrieval, and sparse retrieval. |
Pro/BAAI/bge-m3 |
1,024 |
8,192 |
A multifunctional, multilingual, multi-granularity text embedding model. It supports three common retrieval functions: dense retrieval, multi-vector retrieval, and sparse retrieval. The model can process inputs in over 100 languages and is capable of handling different granularities. |
Variable |
Required |
Description |
|---|---|---|
|
Yes |
Your valid SiliconFlow API key. |