把训练好的深度学习模型优化并部署到 Intel CPU、GPU 等硬件上做推理，支持视觉、LLM 和生成式模型，省去自己折腾底层加速的麻烦。
基于 ONNX Runtime、OpenVINO、PaddlePaddle 等多家推理后端的 OCR 工具包，支持多语言识别，方便在不同硬件上快速部署文字检测与识别。
高性能无锚框 YOLO 目标检测器，精度超过 YOLOv3~v5，支持 MegEngine、ONNX、TensorRT、ncnn 和 OpenVINO 多后端部署。
用小型专用模型搭建企业级 RAG 流水线的统一框架，覆盖文档解析、检索和智能体编排，适合在本地或私有环境部署。
ONNX 是开放的机器学习模型交换格式标准，让 PyTorch、TensorFlow、Keras 等框架训练出的模型能互相转换、跨平台部署。
- Features: - (OpenVINO) Added GroupedMatMul support to the AWQ and Scale Estimation algorithms, enabling data-aware weight compression of grouped mm -based MoE models (h
… the server component and a few improvements from OpenVINO Runtime. Improvements included: - Corrections in input validation for KServer and audio endpoints - Fixed struc
…-Glimmer-30B model: Model card and IR available: [OpenVINO/Muse-Glimmer-30B-int4-ov · Hugging Face](https://huggingface.co/OpenVINO/Muse-Glimmer-30B-int4-ov) Notebooks to
… < 4, including INT2 and INT3 (https://github.com/openvinotoolkit/nncf/pull/4131). - (OpenVINO) Added INT2 and INT3 symmetric compression types (https://github.com/openvi
…mpress weights and nncf.prune (https://github.com/openvinotoolkit/nncf/pull/4063) (https://github.com/openvinotoolkit/nncf/pull/4064). - Fixes: - (PyTorch) Fixed TorchFX
… like a = conv(x); return a a (https://github.com/openvinotoolkit/nncf/pull/3843). - Features: - (OpenVINO) Added NVFP4 ( f4e2m1 ) compression data type in Weight Compres