Skip to content

Deepstream dev vs Deepstream import vision model

Deepstream dev is an open-source agent skill for computer vision. Deepstream import vision model is an open-source agent skill for computer vision.

  • Deepstream dev is best for build a deepstream pipeline that runs object detection on an rtsp camera feed.
  • Deepstream import vision model is best for bring a hugging face rt-detr model into a deepstream pipeline.
What it is
Deepstream dev is an open-source agent skill for computer vision.
Deepstream import vision model is an open-source agent skill for computer vision.
Best for
Build a DeepStream pipeline that runs object detection on an RTSP camera feed
Bring a Hugging Face RT-DETR model into a DeepStream pipeline
Kind
Skill
Skill
Pricing
Open source
Open source
Platforms
Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, Windsurf, Goose
Claude Code, Codex, Cursor, Gemini CLI, GitHub Copilot, Windsurf, Goose
GitHub stars
4K
4K
Last commit
Nov 2026
Nov 2026
License
Apache-2.0
Apache-2.0
Open source
Good for
Build a DeepStream pipeline that runs object detection on an RTSP camera feed; Add a TensorRT-accelerated YOLO model to a video analytics app; Track objects across frames and count them in a video file
Bring a Hugging Face RT-DETR model into a DeepStream pipeline; Export a SafeTensors detector to ONNX and build a TensorRT engine; Benchmark a detection model across several video streams

What Deepstream dev does

  • Minimal pipeline layout

    Describes the standard source, muxer, inference, OSD and renderer chain and which stages are optional.

  • Source handling with nvurisrcbin

    Covers using nvurisrcbin for files, HTTP, HLS and RTSP inputs, including live-source and sync settings.

  • Metadata and buffer rules

    Explains iterator-based metadata access, buffer cloning for async work and request pad naming.

  • Inference configuration

    Covers nvinfer config format, dynamic ONNX dimensions and YOLO output and cluster-mode matching.

  • Kafka message broker integration

    Provides guidance on nvmsgconv and nvmsgbroker placement and their configuration references.

  • Virtual environment setup

    Specifies installing the pyservicemaker wheel and pyyaml inside a Python venv.

What Deepstream import vision model does

  • Model acquisition from hub

    Locates and downloads ONNX files from Hugging Face or NGC, or exports SafeTensors weights to ONNX.

  • TensorRT engine building

    Builds dynamic-batch TensorRT engines and runs trtexec benchmarks at batch size 1 and the maximum batch.

  • Custom bounding-box parser

    Generates an nvinfer parser and configuration for the detection model's output format.

  • Multi-stream DeepStream benchmarks

    Runs single-stream validation and multi-stream sweeps with repeated runs and KITTI detection dumps.

  • Benchmark report generation

    Produces a Markdown, HTML and PDF report with five charts from the collected benchmark data.

  • Containerized execution

    Runs every phase inside the DeepStream container with a shared Python environment, so no host packages are needed.

Deepstream dev vs Deepstream import vision model: questions

Is Deepstream dev or Deepstream import vision model cheaper?

Both are open source. Compare current plans on each official site.

What is the difference between Deepstream dev and Deepstream import vision model?

Deepstream dev is an open-source agent skill for computer vision. Deepstream import vision model is an open-source agent skill for computer vision.

Keep comparing