Model reference · open weights
SFR-Embedding-Mistral is an open-weight embedding model from Salesforce, listed in the AxForge catalogue. AxForge can bring it up on EU-owned hardware for you on request — with the licence handled where one is required.
About
SFR-Embedding by Salesforce Research. The model is trained on top of E5-mistral-7b-instruct and Mistral-7B-v0.1. This project is for research purposes only. Third-party datasets may be subject to additional terms and conditions under their associated licenses. Please refer to specific papers for more details: - MTEB benchmark - Mistral - E5-mistral-7b-instruct More technical details will be updated later. Ethical Considerations This release is for research purposes only in support of an academic paper. Our models, datasets, and code are not specifically designed or evaluated for all downstream purposes. We strongly recommend users evaluate and address potential concerns related to accuracy, safety, and fairness before deploying this model. We encourage users to consider the common limitations of AI, comply with applicable laws, and leverage best practices when selecting use cases, particularly for high-risk scenarios where errors or misuse could significantly impact people’s lives, rights, or safety. For further guidance on use cases, refer to our AUP and AI AUP. How to run Transformers The models can be used as follows: Sentence Transformers MTEB Benchmark Evaluation Check out unilm/e5 to reproduce evaluation results on the BEIR and MTEB benchmark. SFR-Embedding Team (∗indicates lead contributors). Rui Meng Ye Liu Shafiq Rayhan Joty Caiming Xiong Yingbo Zhou Semih Yavuz Citation
Summarised from the published model card. Read the full card on the HuggingFace links below.
Specifications
| Maker | Salesforce |
|---|---|
| Type | Embedding models |
| Parameters (lead) | 7.1B |
| Context | 32k tokens |
| Variants | 1 |
| Runs with | sentence-transformers |
| Released | 2024-01-24 |
| Popularity | 33k downloads / month |
| Likes | 299 |
| Licence | Commercial licence needed |
How it works
Variants
Open weights ship in several sizes and precisions. One page, all the variants — pick the one that fits your GPU. VRAM figures are estimates from model size.
| Variant | Params | Precision | VRAM | Fits 16 GB | Weights |
|---|---|---|---|---|---|
| SFR-Embedding-Mistral | 7.1B | BF16 | ~16.4 GB | ✓ | Weights ↗ |
Benchmarks
As published on the model card — the maker's own numbers, not measured by AxForge.
| Task | Dataset | Metric | Score |
|---|---|---|---|
| Classification | MTEB AmazonCounterfactualClassification (en) | accuracy | 77.925 |
| Classification | MTEB AmazonCounterfactualClassification (en) | ap | 40.868 |
| Classification | MTEB AmazonCounterfactualClassification (en) | f1 | 71.658 |
| Classification | MTEB AmazonPolarityClassification | accuracy | 95.967 |
| Classification | MTEB AmazonPolarityClassification | ap | 94.463 |
| Classification | MTEB AmazonPolarityClassification | f1 | 95.965 |
| Classification | MTEB AmazonReviewsClassification (en) | accuracy | 54.352 |
| Classification | MTEB AmazonReviewsClassification (en) | f1 | 53.637 |
| Retrieval | MTEB ArguAna | ndcg_at_1 | 43.314 |
| Retrieval | MTEB ArguAna | ndcg_at_2 | 54.757 |
| Retrieval | MTEB ArguAna | ndcg_at_3 | 58.847 |
| Retrieval | MTEB ArguAna | ndcg_at_5 | 63.634 |
| Retrieval | MTEB ArguAna | ndcg_at_7 | 65.741 |
| Retrieval | MTEB ArguAna | ndcg_at_10 | 67.171 |
| Retrieval | MTEB ArguAna | ndcg_at_20 | 68.585 |
| Retrieval | MTEB ArguAna | ndcg_at_30 | 68.81 |
| Retrieval | MTEB ArguAna | ndcg_at_50 | 68.932 |
| Retrieval | MTEB ArguAna | ndcg_at_70 | 68.992 |
| Retrieval | MTEB ArguAna | ndcg_at_100 | 69.014 |
| Retrieval | MTEB ArguAna | ndcg_at_200 | 69.014 |
| Retrieval | MTEB ArguAna | ndcg_at_300 | 69.014 |
| Retrieval | MTEB ArguAna | ndcg_at_500 | 69.014 |
| Retrieval | MTEB ArguAna | ndcg_at_700 | 69.014 |
| Retrieval | MTEB ArguAna | ndcg_at_1000 | 69.014 |
Using it via the API
Once AxForge deploys sfr-embedding-mistral for you, it answers on the OpenAI-compatible API — the same base URL and keys as every other model. (sfr-embedding-mistral below is illustrative; you get the exact model name on deployment.)
$ curl -sS https://api.axforge.ai/v1/embeddings \
-H "Authorization: Bearer $AXFORGE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"sfr-embedding-mistral","input":"text to embed"}'
Details
Languages
Tags
Papers
Licence
The weights are open but cc-by-nc-4.0 needs a commercial agreement for business use. AxForge can arrange that licence and host the model for you — you pay AxForge, we settle with the model’s maker. Ask us for a quote. Read the licence ↗
Explore