All skills
deepstream-dev
Build a working DeepStream Python video analytics pipeline with pyservicemaker, from source to inference to output.
- Part of NVIDIA/skills
- 2.5k installs
- Machine Learning · Vision models
When to use it
You have a video file or RTSP stream and a TensorRT model but need a runnable GStreamer pipeline that detects and saves annotated output.
For example
You say
Use deepstream-dev to build a pipeline that reads a local MP4, runs the sample resnet18 detector, draws boxes, and saves an annotated MP4.
You get
- pyservicemaker pipeline script
- nvinfer config file
- annotated MP4 output
How it works
- Adds only the pipeline stages you ask for, such as tracking or Kafka messaging.
- Handles dynamic ONNX models by setting infer-dims so TensorRT engine builds succeed.
- On WSL2 Ubuntu 24 Docker, writes MP4 output instead of using display sinks.
Before installing
Requires an NVIDIA DeepStream SDK install with the pyservicemaker Python package and its reference documents available.
More Vision models skills
See allAlso in this pack
- accelerated-computing-cudf3.2kGuides GPU DataFrame work with cuDF, including pandas acceleration, joins, groupby, file I/O and multi-GPU jobs.
- cuopt-developer2.8kModifies, builds, tests, and debugs NVIDIA cuOpt source code and prepares contributions.
- cuopt-install2.6kInstalls cuOpt for Python, C, or server via pip, conda, or Docker and verifies it works.
- data-designer2.6kGenerates a synthetic dataset from your description of the data you need.