# QIMP SDK release 1.3 Source: [https://docs.qualcomm.com/doc/80-70017-52/topic/qimp_sdk_release_1_3.html](https://docs.qualcomm.com/doc/80-70017-52/topic/qimp_sdk_release_1_3.html) ## Release information Table : Software version | Software | Version | | --- | --- | | Yocto | Kirkstone 4.0.22 | | Kernel | 6.6.52 | Table : Release tag version | Release tag | Version | | --- | --- | | Firmware release tag | r1.0\_00058.0 | | Manifest release tag | qcom-6.6.52-QLI.1.3-Ver.1.1 | | Meta-qcom-extras release tag | r1.0\_00059.0 | | QIMP SDK release tag | qcom-6.6.52-QLI.1.3-Ver.1.1\_qim-product-sdk-1.1.2 | Table : Supported platforms and reference kits | SoC platforms | Reference kits | | --- | --- | | QCS6490 | | | QCS5430 | | | QCS9075 | Qualcomm^®^ IQ9 Beta Evaluation Kit | | QCS8275 | Qualcomm^®^ IQ8 Beta Evaluation Kit | ## Contents of the release The contents of the Qualcomm^®^ Intelligent Multimedia Product (QIMP) SDK release include: - Recipes for building the individual components: - Qualcomm^®^ Intelligent Multimedia SDK (IM SDK) - Lite Runtime (LiteRT) Note: LiteRT was formerly known as TensorFlow Lite. - Qualcomm^®^ Neural Processing SDK - Qualcomm^®^ AI Engine direct SDK - Sample applications that demonstrate how to use the Qualcomm IM SDK to develop AI edge-based applications. To get started with the QIMP SDK, see [Qualcomm Intelligent Multimedia Product (QIMP) SDK Quick Start Guide](https://docs.qualcomm.com/bundle/publicresource/topics/80-70017-51/introduction.html). ## New features The following new features applicable to QCS6490, QCS9075, and QCS8275 are introduced in the QIMP SDK release: - AI/ML supports execution of batched Qualcomm Neural Processing SDK models and batched QNN models. - Container supports application deployment using an API-based approach – GStreamer daemon (GSTD). ## Sample applications | Sample application | Description | Supported SoC | | --- | --- | --- | | **AI/ML applications** | **AI/ML applications** | **AI/ML applications** | | `gst-ai-face-detection` | Collects the live video input from a camera, file, or an RTSP
stream and uses the Qualcomm Neural Network face detection model to
produce a preview with the overlaid AI model output on the HDMI
display. | QCS6490, QCS9075, and QCS8275 | | `gst-ai-face-recognition` | Collects the live video input from a camera or an RTSP stream and
shares this input for face detection and recognition. | QCS6490, QCS9075, and QCS8275 | | **Video application** | **Video application** | **Video application** | | `gst-ai-smartcodec-example` | Reduces the network bandwidth or storage for input feed from a
camera or file source. | QCS6490, QCS9075, and QCS8275 | For the complete list of sample applications supported in the QIMP SDK and instructions on how to run them, see [Sample applications](https://docs.qualcomm.com/bundle/publicresource/topics/80-70017-50/example-applications.html). ## Python sample applications | Sample application | Description | Supported SoCs | | --- | --- | --- | | **AI/ML applications** | **AI/ML applications** | **AI/ML applications** | | `gst-parallel-inference.py` | Supports multiple AI/ML models run in parallel (object detection,
classification, segmentation, and pose detection) on streams from a
camera, file source, or RTSP. | QCS6490, QCS9075, and QCS8275 | | `gst-daisychain-detection-pose.py` | Collects the live video input from a camera or an RTSP stream and
shares this input for face detection and recognition. | QCS6490, QCS9075, and QCS8275 | | **Video application** | **Video application** | **Video application** | | `gst-concurrent-videoplay-composition.py` | Allows concurrent video playback for MP4 AVC (H.264) videos and
performs composition on a video wall display. | QCS6490, QCS9075, and QCS8275 | | **Camera application** | **Camera application** | **Camera application** | | `gst-multi-camera-stream-example.py` | Enables you to stream from two camera sensors
simultaneously. | QCS6490 | For the complete list of Python sample applications supported in the QIMP SDK and instructions on how to run them, see [Python applications](https://docs.qualcomm.com/bundle/publicresource/topics/80-70017-50/python-sample-applications.html). ## Issues resolved - 18 FPS is observed with `gst-ai-parallel-inference` for file source. - `gst-ai-parallel-inference` hangs and black screen is observed at EOS for RTSP source. - RTSP sink streaming fails to play when you use multiple sinks such as RTSP, file sink, and display with two or more input streams in `multi-input-output-object-detection`. - `gst-ai-monodepth` and `gst-ai-parallel-inference` fails on QCS9075. ## Limitations The following are the known limitations in the QIMP SDK release: - When using the `qtioverlay` plugin with detection models, frame drops may occur, especially with many detections. **Solution**: Use `qtivcomposer` for detection-based ML use cases. - A drop of 1–2 FPS may be observed with the three-stream camera use case. - AI/ML parallel inference for 24 streams varies between 25–30 FPS, which is less than the expected 30 FPS. - Segfault occurs while using Ctrl + C for Qualcomm Neural Processing SDK use cases with DLC models. - Frame drops are observed with the QNN plugin while running on GPU delegate. - Low FPS with daisychain detection and pose GStreamer pipeline. - Observe stability issues with gstreamer pipeline in batched model usecases. - High inference time is seen with `deeplabv3_resnet50.dlc`. - Concurrent streams with smart codec functionality encounter pipeline stalls and cleanup errors at EOS. - Pose estimation for the `facemap_3dmm_quantized.bin` (QNN) AI hub model does not work with the Qualcomm IM SDK pipeline due to caps mismatch. - Caps mismatch with the AI hub face detection QNN model. - The `gst-ai-parallel-inference` sample application hangs for QCS8275. - Support for the FastCV engine is disabled in `qtivtransform`. Last Published: Jan 16, 2025 [Next Topic QIMP SDK release 1.2](https://docs.qualcomm.com/bundle/publicresource/80-70017-52/topics/qimp-sdk-release-1-2.md)