bge-m3
Upstream reference: BGE-M3
Publisher: BAAI
BAAI documents BGE-M3 as a multilingual retrieval model that can produce dense, sparse and multi-vector representations.
Publisher-described context, separate from Sigma5C observations and recorded catalog statistics.
Published upstream specifications
- Dense vector dimensions
- 1024 Source
- Published sequence length
- 8192 tokens Source
- Documented languages
- More than 100 Source
- Upstream license label
- MIT Source
Publisher-described uses
- Multilingual dense, sparse and multi-vector text retrieval. Source
Reference limitations
- Retrieval representations are not generated chat answers. Source
A recorded tag and an upstream card association do not verify the deployed checkpoint, quantization, configured context or current hosted availability.
Publisher sources
- BGE-M3 official documentation — BAAI. Checked .
Links
Recorded references do not establish downloadable weights, licensing or current availability.
Where bge-m3 comes from
Company / lab
BAAI
See every model and family from this maker.
Model family
bge-m3
Other versions and sizes in this line.