Voyager SDK

Hello again ! I’m currently exploring the Voyager SDK on my Metis Compute Board and everything works great with the pre-installed models.Now I’d like to integrate my own ONNX model into the SDK.I’ve seen that each model in Voyager uses a corresponding YAML configuration file that defines preprocessing, inference, and postprocessing parameters.Is there any official guide or documentation explaining: how to properly add a custom ONNX model to the SDK, what each section of the YAML file means (e.g., pipeline, models, datasets, operators), especially tricky ones like weight_md5 or else, and any important details or requirements when adapting an external model (like quantization, input format, etc.)? Thanks a lot in advance for your help!If there are any example templates

Hey again !Finally I was able to install the voyager-SDK on my metis compute board ! Turns out I was debugging too much with chat gpt and it really does nonsense things I should have read more thoroughly the user guide for the metis or  the github repository… But now I have the SDK and I can test the different models directly in the SDK :)I successfully ran the yolov5s-v7-coco.yaml via ./inference <modelname> <media> But I can only run this model…I wonder if it’s normal and the other models in the folder ax_models/ are here but not compiled and need a certain action by the user (me), or if it is a bug and I’m doing something wrong ? When I try to run for example the yolov7 model with the argument --no-display just to see if it works : ./inference.py yolov7-coco media/traffic1_1080p.mp4 --no-display I have this error : 2025-11-05 15:46:12.084642319 [W:onnxruntime:Default, device_discov

Hi it’s still me haha,I’ve successfully installed the Voyager SDK v1.4.2 directly on my Metis Compute Board (Voyager Linux 1.0, Ubuntu 22.04 arm64 base).The full installation (install.sh --runtime --development) completed without any critical errors, and all Axelera packages (runtime, compiler, qtools, etc.) appear correctly in the Python environment.However, when I activate the SDK virtual environment and try to run: python3 -m ax_runner --helpi get : No module named ax_runnerHere’s a summary of my setup and verification steps:✅ Environment cat /etc/os-releaselsb_release -a✅ Installation cd ~/voyager-sdkbash install.sh --runtime --developmentsource venv/bin/activateThe install completed with: Installation complete. To activate the environment, type:  source venv/bin/activateWARNING: Failed to refres

Hello,I am using an Axelera board (board type: AXR_BOARD_METIS_OMEGA_PCIE, firmware version: v1.4.2) and woul like to track the power consumption and current consumption of the AIPU while i am using it. I have tried using Axmonitor but it does not display power consumption (as announced in the SDK since i don’t have a 4-Metis AIPU card). Is there a way to monitor current consumption or power consumption (maybe both?), as there is for temperature?

Hi, The ONNX operator support information published on axelera-ai-hub appears to describe the status starting from opset 14 and newer. I would like to confirm the support status for opset 11. Could you please tell me? Best, regardsHiroyuki Kojima

Hi,i’m using the Voyager SDK 1.4 Version. I have compiled an ONNX resnet50 and now i get the error that the libaxruntime.so is missing. I have performed the installation of the sdk according your manual with an username and token.What can I do to build this library?Best regards,Matthias

This one’s a small but tidy release, mainly focused on smoothing out a few crinkles in v1.4.1.🧹 Fixes: The installer no longer messes with system directory permissions (CST-803). Fixed an error that popped up when saving output from multiple streams (SDK-7862). 📚 Docs & Model Updates: ONNX opset docs got a refresh, now covering expanded support for concat and grouped convolutions. Updated performance metrics for the Real-ESRGAN-x4plus model — now reflecting the latest benchmarking results. If you missed v1.4.1, that one brought a handful of bug fixes and a solid performance boost, including double buffering for inference on Raspberry Pi 5 and Portenta X8, which really improves throughput on those platforms.So, if you haven’t updated since the big 1.4 release, this is

Hello,I’m trying to test the AxInferenceNet C++ Integration.When I launch the test with the Voyager SDK’s given code and the yolov8s-coco model, the results of the detection are totally staggered. Display of the result :I followed the AxInferenceNet C++ Integration Tutorial and I’m using Voyager SDK with Ubuntu 22.04.Could someone help me to understand the results and give me a bunch of idea to solve it ? Thanks in advance.

Dear team, Is it possible to use hardware acceleration provided by the chip for encoding into x264 or x265 video stream on gstreamer. I was able to find plugins here but no specific mention for hardware acceleration.Thank you.

I get this error when I try to deploy Real-ESRGAN. Can anyone tell me why?./deploy.py /home/aetina/Desktop/voyager-sdk-release-v1.4/ax_models/zoo/torch/real-esrgan-x4plus-onnx.yaml --num-cal-images 50INFO    : Detected Metis type as m2INFO    : Compiling network real-esrgan-x4plus-onnx /home/aetina/Desktop/voyager-sdk-release-v1.4/ax_models/zoo/torch/real-esrgan-x4plus-onnx.yamlINFO    : Compile model: real-esrgan-x4plus-onnxINFO    : Using representative images from /home/aetina/.cache/axelera/data/coco2017_repr400 with backend ImageReader.PIL,pipeline input color format ColorFormat.RGBINFO    : Prequantizing real-esrgan-x4plus-onnx: real-esrgan-x4plus-onnxINFO    : Quantizing network real-esrgan-x4plus-onnx /home/aetina/Desktop/voyager-sdk-release-v1.4/ax_models/zoo/torch/real-esrgan-x4plus-onnx.yaml real-esrgan-x4plus-onnxINFO    : Compile model: real-esrgan-x4plus-onnxINFO    : Using representative images from /home/aetina/.

Hello, I’m new to voyager SDK and everything related to AXELERA and i need help deploying a custom onnx model that has 2 inputs and 20+ outputs, all i want for now is to manage to run an inference with my own dataset/image and to save to disk the raw outputs of the model. The input is a pair of 2 files. I have data in yuv format and it’s saved in input_y.npy and input_uv.npy. Do you have any idea how can i do this or if it’s possible from the voyager 1.3 or do i need to add a split layer so i need just 1 input and only then i can deploy it?  Or to upgrade the voyager version to 1.4 

Does anyone have any suggestions on how to implement SAHI inference with Yolo via voyager-sdk to run on metis m2?

Our development team requires specialized expertise for Axelera AI platform implementation and edge AI acceleration optimization. We're seeking an experienced consultant with 5-7weeks availability to support us with

Hello everyone,I have been trying to compile the PRL Tracker model (see link below) for about a month using both Hailo and Vitis AI, but so far I have not been able to fully meet the requirements.I am now considering using the Metis Accelerator, but I am not sure whether I would run into the same issues.Has anyone here successfully deployed this model on the Metis Accelerator, or could you confirm if it is supported?Repository: https://github.com/vision4robotics/PRL-Track/tree/main Any guidance or shared experience would be greatly appreciated.Thanks in advance! 

Hello everyone, I am trying to use the Metis PCIe (4GB) in combination with the firefly ITX-3588J.Issue: I can not load the Metis kernel module (driver) into the Linux kernelSetup: I have the Metis PCIe with 4GB of ram  the Firefly ITX-3588J with 16GB of ram with a RockChip. OS: Ubuntu 22.04.4 LTS. I have a Gen3 Pcie with 4 lanes, details below```        LnkCap:    Port #0, Speed 8GT/s, Width x4, ASPM L0s L1, Exit Latency L0s <4us, L1 <16us            ClockPM- Surprise- LLActRep- BwNot- ASPMOptComp+``` So far:I installed the ubuntu OS as described here: https://support.axelera.ai/hc/en-us/articles/25556437653138-System-Imaging-Guide-Firefly-RK3588 I installed the latest 1.

Hi, I'm working on an RPI5 and I have the voyager SDK inside a container with ubuntu 22 and it works. I wanted to try running an inference on a stream using GStreamer. First, I tried to access the stream and I managed with the following command:GST_PLUGIN_PATH=/voyager-sdk/operators/lib AXELERA_DEVICE_DIR=../opt/axelera/device-1.2.5-1/omega LD_LIBRARY_PATH=/voyager-sdk/operators/lib:/opt/axelera/runtime-1.2.5-1/lib:$LD_LIBRARY_PATH gst-launch-1.0 rtspsrc location='rtsp://…..' latency=200 ! rtph264depay ! avdec_h264 ! videoconvert ! autovideosink Now I wanted to try the axinferencenet plugin but I'm having problems. I've tested various commands and formats, such as: GST_PLUGIN_PATH=/voyager-sdk/operators/lib AXELERA_DEVICE_DIR=../opt/axelera/device-1.2.5-1/omega LD_LIBRARY_PATH=/voyager-sdk/operators/lib:/opt/axelera/runtime-1.2.5-1/lib:$LD_LIBRARY_PATH gst-launch-1.0 rtspsrc location='rtsp://...' latency=200 ! rtph264depay ! avdec_h264 ! vide

Hi everyone,I'm currently developing a web application on top of the Axelera PCIe card and Voyager SDK, and I’ve hit a roadblock.Use Case Overview: I want to relay this annotated stream back to my web frontend for live viewing. I’ve tried: Using GStreamer pipelines for HLS (e.g., with hlssink) Embedding video using <video> tag on the frontend pointing to HLS or .mp4 System Details: Hardware: Axelera Metis PCIe Card Software: Voyager SDK (latest) YOLOv8n-based custom model Python 3.10 + GStreamer on Ubuntu 22.04

Hi,Recently I tried running some ROI tracking algorithms (the ones included with opencv, such as CSRT, MOSSE, etc) on an embedded ARM board and found that the tracking performance and latency is not as good as the application needs. One reason is that all these algorithms run on the CPU, which explains why they are not the fastest. Searching around, I found the VOT (Visual object tracking) challenge related to visual object tracking and it seems that most of the state-of-art trackers now use AI/ML. I guess that’s a good thing since in theory it should be easier to run these on dedicated AI accelerator hardware? Anyway, I want to try running a visual tracking algorithm on the Metis hardware to see how fast and well it can work. However, there are so many algorithms and I reckon the Metis may not support all the operators used in those networks? So before I embark on this task, i want

Hi everyone,I’m new to working with the Axelera system. I see that there is a CenterCrop style preprocessing function, but in my use case I need more control. Specifically, I’d like to crop an image (for RTSP streams or training data) by specifying the x-offset, y-offset, width, and height of the region of interest.Is there a built-in preprocessing function in the Axelera SDK/YAML configs for this kind of custom crop? Or is the recommended way to handle this via GStreamer (e.g., videocrop/videobox) before passing frames to axinferencenet?Any help would be greatly appreciated.Thank you! 🙏

Hello.I installed the SDK on Ubuntu 22.04, butit stopped with the following error:[172/194] Install libgstreamer-plugins-bad1.0-dev[172/194] E: Error, pkgProblemResolver::Resolve stopped. This is probably due to a non-modifiable package.ERROR: Failed to install libgstreamer-plugins-bad1.0-devI would appreciate any advice.

Hello,I’m deploying a model onnx (opset 17) and got a large bunch of warnings of unsupported operations. One of them is the unsupported Transpose:WARNING : Configuration of node '/backbone/3/0/Transpose' may not be supported. 'Transpose' parameters:WARNING : perm: [0, 3, 1, 2]WARNING : data: Metadata(shape=(1, 48, 64, 160), is_constant=False)WARNING : transposed: Metadata(shape=(1, 160, 48, 64), is_constant=False)WARNING : Unsatisfied constraint: perm == [0, 1, 2, 3]I guess the rule should have been perm != [0, 1, 2, 3] because here it is requesting a non no-op operation perm = [0, 3, 1, 2].I think this is also causing later the errorERROR : RuntimeError: The expanded size of the tensor (192) must match the existing size (160) at non-singleton dimension 1. Target sizes: [1, 192, 256, 160]. Tensor sizes: [1, 160, 1, 1]as it looks that indeed the transpose has not been appl

Hello everyone,I'm new to the Axelera AI system and am currently working with a Metis PCIe card with 4 GB memory.I want to understand how much of this memory is being used by the neural network model I am deploying. In a way, I'm looking for a tool or command that provides a report similar to what nvidia-smi does for NVIDIA GPUs, but for the Metis core.From my research, I understand that the Metis core's DRAM is a fixed, dedicated resource and that the memory usage is determined at the time of compilation, not at runtime. I've heard that the Voyager SDK's compiler logs might contain this information.My main question is: Where in the compiler output or log files can I find the exact memory footprint and percentage utilization for my model? Is there a specific argument or command I need to use to generate this report?Any guidance on this would be greatly appreciated.Thank you!

The release notes for the v1.4 SDK state;“Compiled models are expected to run on Intel Core-i 12th and 13th generations (x86), AMD Ryzen (x86) and ARM64 host CPUs”.It would be good to support older generation Intel CPUs, so that servers from 5 or more years ago can be repurposed as AI edge accelerators.

Hi,I tried the Yolo11-coco-onnx models from Model Zoo and got terrible results. Can anyone give me some advice? ./inference.py yolo11n-coco-onnx dataset --no-display --show-device-fpsINFO    : Model:      yolo11n-coco-onnx                                         INFO    : Dataset:    CocoDataset-COCO2017INFO    : Date:       2025-08-21 12:33:18.028455INFO    : Inference Time: 67178.89msINFO    : Evaluation Time: 36222.56msINFO    : Evaluation Metrics:INFO    : =========================INFO    : | mAP_box       | 0.03% |INFO    : | mAP50_box     | 0.05% |INFO    : | precision_box | 0.08% |INFO    : | recall_box    | 1.70% |INFO    : =========================INFO    : Key Metric (mAP_box): 0.03%INFO    : Core Temp : 49.0°C                                                    INFO    : Metis : 804.0fpsINFO    : CPU % : 31.4%INFO    : End-to-end : 74.7fps./inf

 SDK v1.3:14:32:45 [INFO] Dump used CLI arguments to: /repo/build/cli_args.json 14:32:45 [INFO] Dump used compiler configuration to: /repo/build/conf.json 14:32:45 [INFO] Input model has static input shape(s): ((1, 1, 2, 2),). Use it for quantization. 14:32:45 [INFO] Data layout of the input model: NCHW 14:32:45 [INFO] Using dataset of size 100 for calibration. 14:32:45 [INFO] In case of compilation failures, turn on 'save_error_artifact' and share the archive with Axelera AI. 14:32:45 [INFO] Quantizing '' using QToolsV2. 14:32:46 [INFO] ONNX model validation can be turned off by setting 'validate_operators' to 'False'. 14:32:46 [INFO] Checking ONNX model compatibility with the constraints of opset 17. Calibrating... ✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨✨ | 100% | 390.54it/s | 100it | 14:32:46 [INFO] Exporting '' using GraphExporterV2. 14:32:47 [INFO] Quantization finished. 14:32:47 [INFO] Quantization took: 1.57 second