Skip to content

[DLStreamer] Updated Release Notes for DLStreamer 2025.2.0 Release #1494

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jmotow merged 4 commits into from
Dec 10, 2025
+474 −390
Merged

[DLStreamer] Updated Release Notes for DLStreamer 2025.2.0 Release #1494

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
5dae1e3
Updated Release Notes for DLStreamer 2025.2.0 Release
jmotow Dec 4, 2025
0a2f420
Editorial corrections based on Release Notes review.
jmotow Dec 5, 2025
2ef34e8
Addressed feedback from review.
jmotow Dec 10, 2025
c8082bb
Merge branch 'main' into RELEASE_NOTES_2025.2
jmotow Dec 10, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
81 changes: 81 additions & 0 deletions libraries/dl-streamer/RELEASE_NOTES.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,86 @@
# Deep Learning Streamer (DL Streamer) Pipeline Framework Release Notes

## Deep Learning Streamer (DL Streamer) Pipeline Framework Release 2025.2.0

Intel® Deep Learning Streamer (Intel® DL Streamer) Pipeline Framework is a streaming media analytics framework, based on GStreamer* multimedia framework, for creating complex media analytics pipelines. It ensures pipeline interoperability and provides optimized media, and inference operations using Intel® Distribution of OpenVINO™ Toolkit Inference Engine backend, across Intel® architecture, CPU, discrete GPU, integrated GPU and NPU.
The complete solution leverages:

- Open source GStreamer\* framework for pipeline management
- GStreamer* plugins for input and output such as media files and real-time streaming from camera or network
- Video decode and encode plugins, either CPU optimized plugins or GPU-accelerated plugins based on VAAPI
- Deep Learning models converted from training frameworks TensorFlow\*, Caffe\* etc.
- The following elements in the Pipeline Framework repository:

| Element | Description |
|---|---|
| [gvadetect](./docs/source/elements/gvadetect.md) | Performs object detection on a full-frame or region of interest (ROI) using object detection models such as YOLOv4-v11, MobileNet SSD, Faster-RCNN etc. Outputs the ROI for detected objects. |
| [gvaclassify](./docs/source/elements/gvaclassify.md) | Performs object classification. Accepts the ROI as an input and outputs classification results with the ROI metadata. |
| [gvainference](./docs/source/elements/gvainference.md) | Runs deep learning inference on a full-frame or ROI using any model with an RGB or BGR input. |
| [gvatrack](./docs/source/elements/gvatrack.md) | Performs object tracking using zero-term, or imageless tracking algorithms. Assigns unique object IDs to the tracked objects. |
| [gvaaudiodetect](./docs/source/elements/gvaaudiodetect.md) | Performs audio event detection using AclNet model. |
| [gvagenai](./docs/source/elements/gvagenai.md) | Performs inference with Vision Language Models using OpenVINO™ GenAI, accepts video and text prompt as an input, and outputs text description. It can be used to generate text summarization from video. |
| [gvaattachroi](./docs/source/elements/gvaattachroi.md) | Adds user-defined regions of interest to perform inference on, instead of full frame. |
| [gvafpscounter](./docs/source/elements/gvafpscounter.md) | Measures frames per second across multiple streams in a single process. |
| [gvametaaggregate](./docs/source/elements/gvametaaggregate.md) | Aggregates inference results from multiple pipeline branches |
| [gvametaconvert](./docs/source/elements/gvametaconvert.md) | Converts the metadata structure to the JSON format. |
| [gvametapublish](./docs/source/elements/gvametapublish.md) | Publishes the JSON metadata to MQTT or Kafka message brokers or files. |
| [gvapython](./docs/source/elements/gvapython.md) | Provides a callback to execute user-defined Python functions on every frame. Can be used for metadata conversion, inference post-processing, and other tasks. |
| [gvarealsense](./docs/source/elements/gvarealsense.md) | Provides integration with Intel RealSense cameras, enabling video and depth stream capture for use in GStreamer pipelines. |
| [gvawatermark](./docs/source/elements/gvawatermark.md) | Overlays the metadata on the video frame to visualize the inference results. |

For the details on supported platforms, please refer to [System Requirements](./get_started/system_requirements.md).
For installing Pipeline Framework with the prebuilt binaries or Docker\* or to build the binaries from the open source, refer to [Intel® DL Streamer Pipeline Framework installation guide](./get_started/install/install_guide_index.md).

### New in this Release
| Element | Description |
|---|---|
| CPU/GPU configurations | Fixed issue with segmentation fault and early exit for testing scenarios with mixed GPU/CPU device combinations. |
| Documentation | Updated documentation for latency tracer. |
| DeepSort | Fixed DeepSORT feature performance issue |
| Motion detection | Fixed issue low performance of motion detection feature |
| NPU/CPU | Fixed issue where NPU inference required inefficient CPU color processing |
| gvawatermark/gvametaconvert/gvaclassify | Fixed memory leaks on Windows OS |
| License Plate Recognition| Fixed sporadic hang on license-plate-recognition sample on GPU on LNL |
| Model-proc | Improved model proc check logic for va backend |
| Video Analytics | Fixed issue with service crashes |
| gvagenai | Enabled memory mapper and add support for prompt-path |
| gvawatermark | Fixed keypoints metadata processing issue |
| gvarealsense | Fixed issue with missed element in dlstreamer image |
|Arc B580 and Flex 170 | Enabled License Plate Recognition Sample on Arc B580 and Flex 170|
| General| Fixed issue for scenario when vacompositor scale-method option didn't take affect|
| Documentation | Fixed bug in the installation guide |
| Models | Fixed warning message "Model quantization runtime does not match." for older models |
| MQTT | Fixed connection to MQTT |
| BMG/NPU | Fixed support for iGPU, BMG and NPU |
| BMG/latency | Fixed issue with scheduling-policy=latency for BMG NX2 |
| gvapython | Fixed issue with same name for many python modules used by gvapython |
| Sample| Fixed issue with draw_face_attributes sample (cpp) on TGL Ubuntu 24|
| Pose estimation | Fixed wrong pose estimation on ARL GPU with yolo11s-pose |
| ARL| Fixed inconsistent timestamp for vehicle_pedestrian_tracking sample on ARL |
| va-surface | Verified if 2nd and 3rd dGPUs (B580) are paired and working on va-surface sharing|
| License Plate Recognition | Fixed application crash |
| ARL | Enabled YoloV11 INT8 for ARL platform |
| model-instance-id | Fixed hangs for pipelines with model-instance-id configured |
| qsvh264dec | Fixed missing element 'qsvh264dec' in Ubuntu24 docker images |
| GETI2.7 | Enabled support for GETI2.7 detection models on LNL |
| GETI2.7 | Enabled support for GETI2.7 va-surface-sharing on GPU on MTL |



### Known Issues

| Issue | Issue Description |
|---|---|
| VAAPI memory with `decodebin` | If you are using `decodebin` in conjunction with `vaapi-surface-sharing` preprocessing backend you should set caps filter using `""video/x-raw(memory:VASurface)""` after `decodebin` to avoid issues with pipeline initialization |
| Artifacts on `sycl_meta_overlay` | Running inference results visualization on GPU via `sycl_meta_overlay` may produce some partially drawn bounding boxes and labels |
| Preview Architecture 2.0 Samples | Preview Arch 2.0 samples have known issues with inference results. |
| Sporadic hang on `vehicle_pedestrian_tracking_20_cpu` sample | Using Tiger Lake CPU to run this sample may lead to sporadic hang at 99.9% of video processing. Rerun the sample as W/A or use GPU instead. |
| Simplified installation process for option 2 via script | In certain configurations, users may encounter visible errors |
| Error when using legacy YoloV5 models: Dynamic resize: Model width dimension shall be static | To avoid the issue, modify `samples/download_public_models.sh` by inserting the following snippet at lines 273 and 280: |
| | python3 - <<EOF ""${MODEL_NAME}""<br>import sys, os<br>from openvino.runtime import Core<br>from openvino.runtime import save_model<br>model_name = sys.argv[1]<br>core = Core()<br>os.rename(f""{model_name}_openvino_model"", f""{model_name}_openvino_modelD"")<br>model = core.read_model(f""{model_name}_openvino_modelD/{model_name}.xml"")<br>model.reshape([-1, 3, 640, 640]) |



## Deep Learning Streamer (DL Streamer) Pipeline Framework Release 2025.1.2

Intel® Deep Learning Streamer (Intel® DL Streamer) Pipeline Framework is a streaming media analytics framework, based on GStreamer* multimedia framework, for creating complex media analytics pipelines. It ensures pipeline interoperability and provides optimized media, and inference operations using Intel® Distribution of OpenVINO™ Toolkit Inference Engine backend, across Intel® architecture, CPU, discrete GPU, integrated GPU and NPU.
Expand Down
Loading