ProductTensorRT RTX
TensorRT RTX
NVIDIA's local AI inference engine optimized for RTX consumer-grade GPUs. It inherits TensorRT's core capabilities including operator fusion, precision calibration, and automatic kernel tuning. It supports C++ development and is suited for deploying local AI applications on desktop platforms.
Timeline (last 90 days)
Oct 1
TensorRT RTX是NVIDIA为RTX系列GPU优化的推理引擎分支，继承了TensorRT的核心优化能力并针对消费级RTX显卡硬件特性做了适配
Unverified50%
Oct 1
TensorRT RTX的典型工作流为：训练框架导出ONNX，TensorRT解析并优化计算图，生成针对目标GPU的序列化引擎文件（.engine/.plan）
Unverified50%