The vLLM Semantic Router project has released 'd3', a 27B multimodal foundation decision model. This model is designed to answer questions based on text, images, and videos, returning probabilities for each answer without generating text. 'd3' shows significant improvement over its predecessor, Decision 2.0, and demonstrates capabilities in handling multiple images and videos within a single request. It also offers fast inference speeds, processing text requests in a median of 55 ms on a single AMD Instinct MI325X. AI
IMPACT Enables faster, more nuanced decision-making in multimodal applications.
RANK_REASON New model release from a known AI lab. [lever_c_demoted from frontier_release: ic=1 ai=1.0]
Read on Hugging Face Trending Models →
- AMD Instinct MI325X
- Decision 2.0
- Decision 3.0
- Google Colab
- Hugging Face Hub
- Kaggle
- transformers
- vLLM Semantic Router
- vllm-sr/d3
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →