E5-Mistral-7B-Instruct
PermissiveMicrosoft · E5 · Released 2024-01
7BEmbedding
A large decoder-based embedding model that topped the MTEB leaderboard at release, trading model size for retrieval quality.
Strengths
- +MIT license, unrestricted use
- +Top-tier retrieval quality on MTEB at release
- +Instruction-conditioned embeddings improve task-specific relevance
Limitations
- -7B parameters make it far slower and more memory-hungry than small embedders like all-MiniLM
- -Overkill for latency-sensitive, high-QPS retrieval
License
MIT
use commercially with attribution niceties
Hardware
runs on a good consumer GPU (or Apple Silicon) with quantization
Links
Stats
intfloat/e5-mistral-7b-instruct— downloads·— likes
via Hugging Face