BAAI / feature-extraction updated 9 months ago

bge-multilingual-gemma2

BGE-Multilingual-Gemma2 is a LLM-based multilingual embedding model. It is trained on a diverse range of languages and tasks based on google/gemma-2-9b. BGE-Multilingual-Gemma2 primarily demonstrates the following advancements: - Diverse training data: The model's training data spans a broad range of languages, including English, Chinese, Japanese, Korean, French, and more.Additionally, the data covers a variety of t...

Params
9.2 B
Context
8,192
Downloads 30d
1.1 M
Likes
203
Commercial use: conditional · gemma Not gated SAFETENSORS View on Hugging Face ↗

Download history

daily snapshots · 10 days
1.2 M1.1 M
Jul 28Jul 31Aug 3Aug 6

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

FileQuantSizeEst. VRAMVerdict on RTX 4090 · 24 GB
model.safetensors f32 37.0 GB 42.5 GB ❌ Won’t fit
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Specifications

Architecture
Gemma2Model
Parameters
9.2 B
Tensor type
F32
Context length
8,192
Vocabulary
256,002
Layers / heads
42 / 16
Licence
gemma
First seen on the Hub
2024-07-25
Training datasets
undisclosed
MTEB NQ (reported)
99.753765932793
MTEB FEVER (reported)
99.259670014621
MTEB ArguAna (reported)
99.644381223329
MTEB MSMARCO (reported)
99.290830945559
MTEB SCIDOCS (reported)
86.365
MTEB SciFact (reported)
99.666666666667
MTEB FiQA2018 (reported)
97.830697251994
MTEB HotpotQA (reported)
98.399729912221
MTEB NFCorpus (reported)
68.317715094048
MTEB TRECCOVID (reported)
50.920584329202
MTEB ClimateFEVER (reported)
89.415852334419
MTEB QuoraRetrieval (reported)
90.037603230061
Added to our catalog
2026-07-28