Video streaming inference framework, integrating image algorithms and models for real-time/offline video structuring, lightweight NVIDIA DeepStream. 视频流式推理框架,结合图像算法和模型进行实时/离线视频结构化,轻量版 NVIDIA DeepStream
Yes, because video analytics understanding tasks often require distributed computing power, stream-infer is designed to support only a single stream.
But you can still write a simple task queue scheduling application, enable multiple stream-infer docker containers, or use a multi-process scheduling framework like dynamic_app.py and dynamic_task_schedule.py
Yes, because video analytics understanding tasks often require distributed computing power, stream-infer is designed to support only a single stream. But you can still write a simple task queue scheduling application, enable multiple stream-infer docker containers, or use a multi-process scheduling framework like dynamic_app.py and dynamic_task_schedule.py