NVIDIA Releases Multilingual Embedding Models
tomaarsen · x · 2026-07-17
NVIDIA has released a new embedding solution natively integrated into Sentence Transformers. Users can directly load nvidia/Nemotron-3-Embed-8B-BF16 to encode queries and documents separately.
The post notes support for an NVFP4 variant and vLLM. It preserves prompts and normalization metadata while automatically handling query/passage prefixes. The author considers this a major release for multilingual RAG, showing leading retrieval accuracy, and suggests the 1B version is better suited for budget-conscious teams.
Related event: NVIDIA launches Nemotron-3-Embed retrieval models(11 posts)→
More from Infra
- AWS on agent prompt caching: explicit + implicit patterns to slash token costs — AWS ML Blog · 2026-10-27
- Thieves spoofing GPS steal $100M+ in data center cargo since June — Ars Technica AI · 2026-10-27
- Lam Research: weekly token usage up ~250x in 18 months — Beth_Kindig · 2026-10-27
- AWS shows how to build enterprise agentic retrieval on Bedrock with CDK — AWS ML Blog · 2026-10-27
- Encrypted GPT-2 speedrun challenges FHE inference: best is 3.031 s/token, goal sub-1s — tensorqt · 2026-10-27
- Open-source Latch compiles PyTorch into training megakernels, 2.75x faster on H100 — sguada · 2026-10-27