[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-93530":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":9,"totalLinesOfCode":9,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":9,"subscribersCount":16,"size":16,"stars1d":16,"stars7d":17,"stars30d":17,"stars90d":16,"forks30d":16,"starsTrendScore":17,"compositeScore":18,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":19,"hasPages":19,"topics":9,"createdAt":9,"pushedAt":9,"updatedAt":21,"readmeContent":22,"aiSummary":9,"trendingCount":16,"starSnapshotCount":16,"syncStatus":23,"lastSyncTime":24,"discoverSource":25},93530,"DeepStream","NVIDIA\u002FDeepStream","NVIDIA","NVIDIA DeepStream Monorepo: DeepStream SDK and reference apps for building GPU‑accelerated, real-time video and multi‑sensor analytics pipelines with GStreamer, TensorRT, and vision AI models across edge, on‑prem, and cloud deployments.",null,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDeepStream","C++",91,19,3,6,0,5,43.9,false,"main","2026-07-22 04:02:09","# DeepStream\n\n[NVIDIA DeepStream SDK](https:\u002F\u002Fdeveloper.nvidia.com\u002Fdeepstream-sdk) is a streaming analytics toolkit for AI-based video and image understanding, providing a GStreamer-based framework to build multi-stream, multi-model inference pipelines on NVIDIA GPUs (dGPU and Jetson).\n\nDeepStream pipelines combine hardware-accelerated decoding\u002Fencoding, TensorRT inference, object tracking, and message-broker integrations to deliver real-time video analytics across dGPU and Jetson platforms.\n\nDeepStream is published as a unified GitHub repository; sources are to be used under the project [LICENSE](LICENSE) (documentation: CC-BY-4.0; source code: Apache-2.0). GitHub Release assets specific license details can be referred at [Release Asset License](#release-asset-license) under [DeepStream Release Assets](#deepstream-release-assets). The project is actively maintained, but currently not accepting code contributions — refer to the [Contribution Guidelines](#contribution-guidelines) for more details.\n\n# Overview\n\nThis repository contains the complete source code for DeepStream 9.1.\n\n**Components** ([`src\u002F`](src\u002FREADME.md)):\n- [`src\u002Fgst-plugins\u002F`](src\u002FREADME.md#gstreamer-plugins) — DeepStream GStreamer plugin sources\n- [`src\u002Futils\u002F`](src\u002FREADME.md#utility-libraries) — utility library sources\n- [`src\u002Fapps\u002Fsample_apps\u002F`](src\u002FREADME.md#sample-applications) — GStreamer-based sample applications\n- [`src\u002Fapps\u002Freference_apps\u002F`](src\u002FREADME.md#reference-applications) — advanced reference applications\n- [`src\u002Fapps\u002Ftao_apps\u002F`](src\u002FREADME.md#tao-apps) — TAO-model integration apps\n- [`src\u002Fservice-maker\u002F`](src\u002FREADME.md#service-maker) — Service Maker C++\u002FPython SDK and apps\n\n**Tools** ([`tools\u002F`](tools\u002F)):\n- [`inference_builder`](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Finference_builder\u002Ftree\u002F3f0c09f2e3da076cbbb75e17bdebd565b03d1a18) — visual inference pipeline builder\n- [`auto-magic-calib`](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Fauto-magic-calib\u002Ftree\u002Fbc5b1d9b3154d4242f147c35754670bc7981d37a) — camera auto-calibration tool\n- [`yolo_deepstream`](tools\u002Fyolo_deepstream\u002FREADME.md) — YOLO + TensorRT integration\n- [`sam2-onnx-tensorrt`](tools\u002Fsam2-onnx-tensorrt\u002FREADME.md) — SAM2 ONNX-to-TensorRT conversion\n- [`deepstream-host-provisioning`](tools\u002Fdeepstream-host-provisioning\u002FREADME.md) — Ansible playbooks to provision Ubuntu hosts (x86_64 & ARM SBSA) with the full NVIDIA compute stack (driver, CUDA, cuDNN, TensorRT, Docker) and DeepStream dependencies\n- [`deepstream-otelcol`](tools\u002Fdeepstream-otelcol\u002FREADME.md) — OpenTelemetry Collector setup for streaming metrics\n\n**AI agent skills** ([`skills\u002F`](skills\u002FREADME.md), for Claude Code & compatible coding agents):\n- [`deepstream-dev`](skills\u002Fdeepstream-dev\u002FSKILL.md) — general DeepStream development\n- [`deepstream-generate-pipeline`](skills\u002Fdeepstream-generate-pipeline\u002FSKILL.md) — interactive `gst-launch` pipeline builder\n- [`deepstream-profile-pipeline`](skills\u002Fdeepstream-profile-pipeline\u002FSKILL.md) — Nsight Systems profiling & config derivation\n- [`deepstream-sop`](skills\u002Fdeepstream-sop\u002FSKILL.md) — SOP step-sequence compliance microservice\n- [`deepstream-import-vision-model`](skills\u002Fdeepstream-import-vision-model\u002FSKILL.md) — autonomous vision-model onboarding\n- [`deepstream-run-mv3dt`](skills\u002Fdeepstream-run-mv3dt\u002FSKILL.md) — run the MV3DT reference app on sample or custom datasets\n- [`amc-setup-calibration-stack`](skills\u002Famc-setup-calibration-stack\u002FSKILL.md) — launch the AutoMagicCalib microservice and UI\n- [`amc-run-sample-calibration`](skills\u002Famc-run-sample-calibration\u002FSKILL.md) — verify AMC with the bundled sample dataset\n- [`amc-run-video-calibration`](skills\u002Famc-run-video-calibration\u002FSKILL.md) — calibrate user-provided MP4 camera videos\n- [`amc-run-rtsp-calibration`](skills\u002Famc-run-rtsp-calibration\u002FSKILL.md) — calibrate live RTSP camera streams through VIOS capture\n\n# Requirements\n\nBefore building, ensure the following prerequisites are installed:\n\n- **NVIDIA compute stack** — driver, CUDA, cuDNN, and TensorRT at the versions listed below. See the [DeepStream SDK Installation Guide](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_Installation.html).\n- **DeepStream 9.1 proprietary runtime** — the prebuilt libraries and sample data are published as GitHub Release assets. `bash build\u002Fbuild.sh` automatically downloads the packages into `artifacts\u002F` before the build process starts.\n\n  > **Note:** After a successful build, `artifacts\u002F` is removed automatically. Pass `--keep-assets` to retain them.\n\n> **SBSA \u002F DGX Spark:** DeepStream bare-metal installation is not supported on SBSA. Build and validation must be done inside the NVIDIA SBSA Docker container, which ships with DeepStream pre-installed — `build\u002Fbuild.sh` is **not required** inside the container. If you modify an open-source component, run `make && make install` in that component's directory to deploy the updated binaries\u002Flibs. See [build\u002FBUILD.md](build\u002FBUILD.md).\n\n# Getting Started\n\n```bash\n# Install Git LFS (required for LFS-tracked sample media in reference apps)\nsudo apt-get install git-lfs && git lfs install\n\n# Clone the repo with submodules\ngit clone --recurse-submodules https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDeepStream.git && cd DeepStream\n```\n\nSee **[build\u002FBUILD.md](build\u002FBUILD.md)** for full build instructions, including system package dependencies (x86, aarch64, SBSA \u002F DGX Spark), `build\u002Fbuild.sh` usage and environment variables (`CUDA_VER`, `NVDS_VERSION`, `INSTALL_METHOD`), and build output locations under `\u002Fopt\u002Fnvidia\u002Fdeepstream\u002Fdeepstream-9.1\u002F`.\n\n**Supported Platforms**\n\n| Platform | Architecture | Notes |\n|---|---|---|\n| x86 dGPU | x86_64 | Ubuntu 24.04, CUDA 13.2, TensorRT 10.16.x, driver 595+ |\n| Jetson | aarch64 | JetPack 7.2 GA (CUDA 13.2, TensorRT 10.16.x) |\n| SBSA \u002F DGX Spark | aarch64 | Bare-metal DS installation not supported; build and run inside the NVIDIA SBSA Docker container. Only open-source components are built — no artifacts installed. |\n\n```bash\n# Build. The script prompts for sudo only when installing to system paths.\nbash build\u002Fbuild.sh\n```\n\n# Usage\n\nAfter `bash build\u002Fbuild.sh`, binaries are installed to `\u002Fopt\u002Fnvidia\u002Fdeepstream\u002Fdeepstream-9.1\u002Fbin\u002F`. Run the reference `deepstream-app` with one of the sample configs:\n\n```bash\ncd \u002Fopt\u002Fnvidia\u002Fdeepstream\u002Fdeepstream-9.1\u002Fsamples\u002Fconfigs\u002Fdeepstream-app\ndeepstream-app -c source30_1080p_dec_infer-resnet_tiled_display.txt\n\n# After the first install, clear the GStreamer plugin cache if needed:\nrm -rf ~\u002F.cache\u002Fgstreamer-1.0\u002F\n```\n\n- More such examples: [DeepStream Quickstart Guide](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_Quickstart.html#run-deepstream-app-the-reference-application)\n- Detailed API reference: [DeepStream SDK Developer Guide](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002F)\n\nEach app must be run from its source directory so relative config paths resolve correctly. Refer to the `README` inside each app directory for app-specific run instructions and config options.\n\n## Running with Triton Inference Server (Docker)\n\n> **Optional.** Skip this if the bare-metal build above already works for you. Use the Triton container when you need Triton-backed inference. On SBSA \u002F DGX Spark, use the SBSA Docker container as described above — bare-metal installation is not supported.\n\nNVIDIA publishes a DeepStream Docker image bundled with Triton Inference Server.\n\n```bash\n# One-time NGC login (get an API key from https:\u002F\u002Fngc.nvidia.com)\ndocker login nvcr.io           # username: $oauthtoken,  password: \u003CNGC API key>\n\n# Pull the image (use 9.1-triton-arm-sbsa instead on SBSA \u002F DGX Spark)\ndocker pull nvcr.io\u002Fnvidia\u002Fdeepstream:9.1-triton-multiarch\n\n# Launch with display (use 'fakesink' in your pipeline for headless)\nexport DISPLAY=:0 && xhost +\ndocker run -it --rm --gpus all --network=host \\\n    -e DISPLAY=$DISPLAY -v \u002Ftmp\u002F.X11-unix\u002F:\u002Ftmp\u002F.X11-unix \\\n    nvcr.io\u002Fnvidia\u002Fdeepstream:9.1-triton-multiarch\n\n# Inside the container, run a Triton-backed sample\ncd \u002Fopt\u002Fnvidia\u002Fdeepstream\u002Fdeepstream-9.1\u002Fsamples\u002Fconfigs\u002Fdeepstream-app-triton\ndeepstream-app -c source30_1080p_dec_infer-resnet_tiled_display.txt\n```\n\n**Prerequisites:** Docker (`docker-ce`), the [NVIDIA Container Toolkit](https:\u002F\u002Fdocs.nvidia.com\u002Fdatacenter\u002Fcloud-native\u002Fcontainer-toolkit\u002Finstall-guide.html), NVIDIA driver 595+, and an NGC API key. Triton sample model repos ship under `\u002Fopt\u002Fnvidia\u002Fdeepstream\u002Fdeepstream\u002Fsamples\u002Ftriton_model_repo\u002F`. For gRPC-backed Triton, use `samples\u002Fconfigs\u002Fdeepstream-app-triton-grpc\u002F` instead.\n\n# Documentation\n\n| Page | Description |\n|------|-------------|\n| [Overview & Architecture](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_Overview.html) | What DeepStream is, key features, and the GStreamer-based pipeline architecture. |\n| [Release Notes](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_Release_notes.html) | What's new in this release, supported platforms, and known issues. |\n| [Installation](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_Installation.html) | Step-by-step installation of the NVIDIA compute stack and DeepStream SDK on x86 and Jetson. |\n| [Docker Containers](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_docker_containers.html) | Available DeepStream NGC container images (incl. Triton variants) and how to pull and run them. |\n| [DeepStream Samples](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_C_Sample_Apps.html) | Reference walkthroughs of the bundled C\u002FC++ sample applications and what each one demonstrates. |\n| [Reference Applications](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_ref_app_github.html) | Advanced GitHub-hosted reference apps demonstrating specialized end-to-end pipelines. |\n| [TAO Apps](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_TAO_integration.html) | DeepStream sample apps integrating TAO-trained models (detection, classification, segmentation, pose, LPR). |\n| [DeepStream Plugins](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_plugin_Intro.html) | Reference for every DeepStream GStreamer plugin — properties, pad caps, and usage. |\n| [Service Maker](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_service_maker_intro.html) | Build DeepStream pipelines declaratively with the C++ \u002F Python Service Maker SDK. |\n| [Inference Builder](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Finference_builder\u002Ftree\u002F3f0c09f2e3da076cbbb75e17bdebd565b03d1a18) | Compose and configure DeepStream inference pipelines visually with the Inference Builder tool. |\n| [Auto Magic Calib](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Fauto-magic-calib\u002Ftree\u002Fbc5b1d9b3154d4242f147c35754670bc7981d37a) | Automatic camera calibration tool for multi-camera DeepStream deployments. |\n| [DeepStream Libraries](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_Libraries.html) | DeepStream utility and helper libraries bundled as the `deepstream_libraries` submodule. |\n| [DeepStream Coding Agent](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_AI_Agent.html) | Use the bundled `skills\u002F` with Claude Code and other AI coding assistants to generate DeepStream pipelines. |\n\n# Repository Structure\n\n```\nDeepStream\u002F\n├── .claude\u002F                                 # Claude Code project-local config (settings, commands)\n├── .github\u002F                                 # GitHub workflows, issue templates, CODEOWNERS\n├── build\u002F\n│   ├── BUILD.md                             # build instructions\n│   └── build.sh                             # top-level build driver\n├── example_prompts\u002F                         # example prompts for AI coding agents\n├── includes\u002F                                # shared public headers (ds3d, nvdsinferserver, …)\n├── scripts\u002F\n│   ├── install_artifacts.sh                 # installs proprietary libs + sample data to \u002Fopt\n│   ├── install_opensource_deps.sh           # builds open-source deps (OpenTelemetry, civetweb, …)\n│   ├── install.sh                           # system integration (update-alternatives, pyservicemaker)\n│   ├── uninstall.sh                         # removes a DeepStream installation from \u002Fopt\n│   ├── triton_backend_setup.sh              # sets up Triton inference server backends\n│   ├── prepare_ds_triton_model_repo.sh      # prepares the Triton model repository\n│   ├── prepare_ds_triton_tao_model_repo.sh  # prepares the Triton model repository for TAO models\n│   ├── prepare_classification_test_video.sh # generates a classification test video\n│   ├── rtpmanager\u002Fupdate_rtpmanager.sh      # rtpjitterbuffer\u002Frtsp EOS-hang fix script + patch files\n│   └── print_env.sh                         # env diagnostic helper for bug reports\n├── skills\u002F\n│   ├── amc-run-sample-calibration\u002F          # AMC sample-dataset calibration skill\n│   ├── amc-run-video-calibration\u002F           # AMC custom-video calibration skill\n│   ├── amc-run-rtsp-calibration\u002F            # AMC RTSP-stream calibration skill\n│   ├── amc-setup-calibration-stack\u002F         # AMC microservice + UI launch skill\n│   ├── deepstream-dev\u002F                      # general DS development skill\n│   ├── deepstream-generate-pipeline\u002F        # interactive gst-launch pipeline builder\n│   ├── deepstream-profile-pipeline\u002F         # Nsight Systems profiling & config derivation\n│   ├── deepstream-sop\u002F                      # SOP step-sequence compliance microservice\n│   ├── deepstream-import-vision-model\u002F      # autonomous vision-model onboarding skill\n│   └── deepstream-run-mv3dt\u002F                # MV3DT reference-app operations skill\n├── src\u002F\n│   ├── apps\u002F                                # sample, reference, and TAO sample applications\n│   ├── gst-plugins\u002F                         # GStreamer plugin sources (per-plugin subdirs)\n│   ├── gst-utils\u002F                           # GStreamer utility library sources\n│   ├── service-maker\u002F                       # Service Maker C++\u002FPython SDK and apps\n│   └── utils\u002F                               # utility library sources (per-library subdirs)\n├── deepstream_libraries\u002F                    # DeepStream utility and helper libraries (submodule)\n├── deepstream_dockers\u002F                      # DeepStream dockers\n└── tools\u002F\n    ├── auto-magic-calib\u002F                    # camera auto-calibration tool\n    ├── deepstream-otelcol\u002F                  # OpenTelemetry Collector setup for streaming metrics\n    ├── inference_builder\u002F                   # visual inference pipeline builder\n    ├── deepstream-host-provisioning\u002F        # Ansible playbooks for host provisioning (x86_64 & ARM SBSA)\n    ├── sam2-onnx-tensorrt\u002F                  # SAM2 ONNX-to-TensorRT conversion\n    └── yolo_deepstream\u002F                     # YOLO + TensorRT integration\n```\n\n> **NOTE:** DeepStream-specific repositories previously hosted under the NVIDIA-AI-IOT GitHub organization are now consolidated into this repository.\n\nThe table below maps the existing NVIDIA-AI-IOT GitHub repository sources to their location in this repository.\n\n\n| NVIDIA-AI-IOT GitHub repo                                                               | Location in this repository                            |\n| --------------------------------------------------------------------------------------- | ------------------------------------------------------ |\n| [deepstream_tao_apps](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Fdeepstream_tao_apps)             | [src\u002Fapps\u002Ftao_apps\u002F](src\u002Fapps\u002Ftao_apps\u002F)               |\n| [deepstream_reference_apps](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Fdeepstream_reference_apps) | [src\u002Fapps\u002Freference_apps\u002F](src\u002Fapps\u002Freference_apps\u002F)   |\n| [deepstream_tools](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Fdeepstream_tools)                   | [tools\u002F](tools\u002F)                                       |\n| [DeepStream_Coding_Agent](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002FDeepStream_Coding_Agent)     | [skills\u002F](skills\u002FREADME.md)                            |\n| [inference_builder](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Finference_builder)                 | [tools\u002Finference_builder\u002F](tools\u002Finference_builder\u002F)   |\n| [auto-magic-calib](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Fauto-magic-calib)                   | [tools\u002Fauto-magic-calib\u002F](tools\u002Fauto-magic-calib\u002F)     |\n| [deepstream_dockers](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Fdeepstream_dockers)               | [deepstream_dockers\u002F](deepstream_dockers\u002F)             |\n| [deepstream_libraries](https:\u002F\u002Fgithub.com\u002FNVIDIA-AI-IOT\u002Fdeepstream_libraries)           | [deepstream_libraries\u002F](deepstream_libraries\u002F)         |\n\n\n\n# Performance\nSummary of benchmarks; for detailed performance numbers and hardware used, refer to the [DeepStream Performance Guide](https:\u002F\u002Fdocs.nvidia.com\u002Fmetropolis\u002Fdeepstream\u002Fdev-guide\u002Ftext\u002FDS_Performance.html).\n\n# Releases & Roadmap\n\n- **Releases \u002F Changelog:** [CHANGELOG.md](.\u002FCHANGELOG.md)\n\n## DeepStream Release Assets\n\nThe following `.deb` and `.tar` artifacts are published as GitHub Release assets. `build\u002Fbuild.sh` downloads them automatically into `artifacts\u002F` before installation:\n\n### Debian packages (`.deb`)\n\n| Asset | Architecture | Description |\n|---|---|---|\n| `deepstream-9.1_9.1.0-1_amd64.deb` | x86_64 | DeepStream 9.1 SDK Debian package for x86 dGPU. |\n| `deepstream-9.1_9.1.0-1_arm64.deb` | aarch64 | DeepStream 9.1 SDK Debian package for Jetson. |\n| `deepstream-binaries-x86_9.1.0_amd64.deb` | x86_64 | Prebuilt proprietary runtime libraries for x86. |\n| `deepstream-binaries-aarch64_9.1.0_arm64.deb` | aarch64 | Prebuilt proprietary runtime libraries for Jetson. |\n| `deepstream-sample-data_9.1.0.deb` | all | Sample models, configs, and video streams. |\n\n### Tarball archives (`.tar.gz` \u002F `.tbz2`)\n\n| Asset | Architecture | Description |\n|---|---|---|\n| `deepstream_sdk_v9.1.0_x86_64.tbz2` | x86_64 | DeepStream 9.1 SDK tarball for x86 dGPU. |\n| `deepstream_sdk_v9.1.0_jetson.tbz2` | aarch64 | DeepStream 9.1 SDK tarball for Jetson. |\n| `deepstream-binaries-x86_9.1.0.tar.gz` | x86_64 | Prebuilt proprietary runtime libraries for x86. |\n| `deepstream-binaries-aarch64_9.1.0.tar.gz` | aarch64 | Prebuilt proprietary runtime libraries for Jetson. |\n| `deepstream-sample-data_9.1.0.tar.gz` | all | Sample models, configs, and video streams. |\n\n### Python wheels (`.whl`)\n\n| Asset | Architecture | Description |\n|---|---|---|\n| `deepstream_libraries-1.4-cp312-cp312-linux_x86_64.whl` | x86_64 | DeepStream Libraries Python wheel (CPython 3.12) for x86. |\n\n### Download installer(s)\n\nRelease assets can be downloaded via `curl` or `wget` from the command line.\n\n**curl**\n\n```bash\ncurl -LO 'https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDeepStream\u002Freleases\u002Fdownload\u002Fv9.1.0\u002Fdeepstream-9.1_9.1.0-1_amd64.deb'\n```\n\n**wget**\n\n```bash\nwget --content-disposition 'https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDeepStream\u002Freleases\u002Fdownload\u002Fv9.1.0\u002Fdeepstream-9.1_9.1.0-1_amd64.deb'\n```\n\n> **NOTE:** Replace `deepstream-9.1_9.1.0-1_amd64.deb` with the name of the asset you want to download.\n\n## Docker Images (NGC)\n\nDeepStream Docker images are released through NGC. Browse and pull the images from the [DeepStream NGC container catalog](https:\u002F\u002Fcatalog.ngc.nvidia.com\u002Forgs\u002Fnvidia\u002F-\u002Fcontainers\u002Fdeepstream\u002F-?_lr=1).\n\n> **NOTE:** For `docker login` and `docker pull` command details, along with the full list of available image tags, refer to the [DeepStream NGC container catalog](https:\u002F\u002Fcatalog.ngc.nvidia.com\u002Forgs\u002Fnvidia\u002F-\u002Fcontainers\u002Fdeepstream\u002F-?_lr=1).\n\n# Contribution Guidelines\nThis project is currently not accepting contributions. The product roadmap is managed internally by NVIDIA.\n\n## Security\n- Vulnerability disclosure: File an issue at [Issues](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDeepStream\u002Fissues).\n- Do not file public issues for security reports.\n\n## Support\n- Level: Maintained\n- How to get help: [Issues](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDeepStream\u002Fissues) \u002F [DeepStream SDK Developer Forum](https:\u002F\u002Fforums.developer.nvidia.com\u002Fc\u002Faccelerated-computing\u002Fintelligent-video-analytics\u002Fdeepstream-sdk\u002F15)\n- Filing a bug? Use the [bug report template](.github\u002FISSUE_TEMPLATE\u002Fbug_report.md) and include: platform (x86 \u002F Jetson \u002F SBSA), DeepStream version, install method (Debian \u002F Docker \u002F source), and the output of `scripts\u002Fprint_env.sh`.\n\n# Community\n\nJoin the DeepStream community to ask questions, share feedback, and report issues.\n\n- [GitHub Issues](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDeepStream\u002Fissues)\n- [GitHub Discussions](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FDeepStream\u002Fdiscussions)\n- [DeepStream SDK Developer Forum](https:\u002F\u002Fforums.developer.nvidia.com\u002Fc\u002Faccelerated-computing\u002Fintelligent-video-analytics\u002Fdeepstream-sdk\u002F15)\n\n# References\n\n- [DeepStream NGC container](https:\u002F\u002Fcatalog.ngc.nvidia.com\u002Forgs\u002Fnvidia\u002Fcontainers\u002Fdeepstream) — prebuilt runtime image (also bundles Triton).\n- [NVIDIA TAO Toolkit](https:\u002F\u002Fdeveloper.nvidia.com\u002Ftao-toolkit) — model training\u002Foptimization, source of the TAO sample apps.\n- [GStreamer documentation](https:\u002F\u002Fgstreamer.freedesktop.org\u002Fdocumentation\u002F) — underlying framework.\n\n# License\nThis project is licensed under [CC-BY-4.0 AND Apache-2.0](.\u002FLICENSE).\n- SPDX-License-Identifier: CC-BY-4.0 AND Apache-2.0\n\n### Release Asset License\n\nAll GitHub Release assets (`.deb` packages and `.tar.gz` tarballs under the `deepstream-binaries-*` and `deepstream-sample-data_*` names) are governed by the [Software License Agreement For NVIDIA Software Development Kits](https:\u002F\u002Fdeveloper.download.nvidia.com\u002Fassets\u002FDeepstream\u002FDeepStream_6.0\u002FNVIDIA_DeepStream_SDK_EULA_6.0.pdf), refer the \"LicenseAgreement\" included within all GitHub Release Assets packages.\n\n## Third‑Party License\u002FNotice\n\nRefer [THIRD_PARTY_LICENSES](.\u002FTHIRD_PARTY_LICENSES.txt) for all 3rd Party OSS licenses and Notices\n\n",2,"2026-07-20 02:30:05","trending"]