The Axelera Forum

Thought you might want to check out the live stream from YOLO Vision that’s happening today in London. Some of the Axelera AI team are there, and will be up on stage soon! 

Hello,I am currently running performance tests on the Metis M.2 card with different models (e.g., YOLOv8n).I observed a behavior that I would like to clarify: When running with 1 AIPU core, I get lower latency and the same FPS as when using multiple cores (4). However, with 4 AIPU cores, the latency per image is higher, even though the FPS does not increase. In addition, CPU utilization is noticeably higher with 4 cores (around 17%) compared to only ~0.8% with 1 core. I would like to know: Is this expected behavior (due to synchronization overhead between AIPU cores)? Is there a recommended configuration to optimize either latency or throughput (FPS) depending on the use case? Is it normal that CPU usage increases significantly when using 4 cores, given that processing is supposed to be largely handled by the AIPU? Thank you in advance for your clarifications.Best regards,

In case anyone on here hasn’t spotted it in the wild yet, the Metis M.2 Max was just announced. It’s basically managed to pack PCIe card-level performance into an M.2 stick. So that’s better performance for LLMs and vision transformers on much smaller host devices (RPi, Orange Pi, etc), all while keeping power draw super low (around 6.5W). It’s been bumped up to 16GB of memory, slimmer design options, and added security features for tougher environments.Feels like a big step forward for anyone wanting to run GenAI, computer vision, or LLM workloads locally, on devices with an impressively small footprint, without needing a data centre to do it 😆.Shipping starts later this year.https://axelera.ai/news/axelera-ai-boosts-llms-at-the-edge-by-2x-with-metis-m.2-max-introduction

Heads-up if you’re building a new AI project from scratch – the new Metis Compute Board is now up for pre-order!It’s a compact SBC with the Metis AI accelerator baked right in, powered by the ARM-based RK3588. And it runs the Voyager SDK out of the box. Loads of interfaces, solid memory specs, tiny footprint (170x170mm), and very dev-friendly.Integrated Metis AIPU with 214 TOPS for multi-stream, multi-model AI workloads 16 GB RAM for the CPU + 4 GB for the AIPU, with storage and connectivity for almost any setup Comes with Voyager SDK and BSP for fast prototyping and deploymentThere’s a product brief attached if you want to see a few more details, too.Bonus: pre-orders get 15% off. Shipping is slated to begin in June 2025, and you get 15% off for pre-ordering.Check it out: https://store.axelera.ai/products/metis-compute-board-with-arm-based-rk3588 

All the release notes for Voyager updates go straight out on the GitHub repo, so I don’t tend to parrot them here, but this is a pretty big one and feels like it deserves a spot here on the community!If you’ve been using 1.3 already, there are some nice changes that make life easier.A few things that particularlt stand out to me:🔹 Cascade pipelines feel smoother. Chaining models together (like running a classifier on detections) takes less fiddling around than before. 🔹 More operator and metadata options. You can now shape pre/post-processing and inference results in ways that fit better with your own app. 🔹 Model Zoo is easier to adapt. Swapping in your own weights on the reference pipelines is less hassle, so you can get something tuned to your dataset faster. 🔹 Custom evaluators are handy if you want accuracy measured in a way that matches your own project rather than sticking with defaults. 🔹 General quality-of-life polish in pipeline debugging, deployment consistency, benchma

We are happy to announce the release of our latest PySDK (version 0.18.0) that supports Axelera runtime 1.3.3. See our docs: Overview | DeGirum Docs for more details. Examples and user guides coming soon.

I suppose it’s inevitable that there’ll always be a new model in the works, but hushed conversations about the Orange Pi 6 Plus are starting to surface…Considering how well the Orange Pi 5 Plus has (rather unexpectedly) performed in low-cost edge AI applications, it’s hard not to feel some small pang of excitement around the next version!https://www.reddit.com/r/OrangePI/comments/1mbd5zc/the_upcoming_orangepi_6_plus/We need to put one in the hands of ​@Victor Labian and ​@Steven Hunsche! 

We're excited to announce that Axelera is now integrated with DeGirum AI Hub—a platform that lets you run edge AI models directly in your browser without any setup or installations.🔍 What You Can DoExplore a wide range of ready-to-use AI models for real-time inference across various use cases: 🧑 Face Detection – detect and track faces in video streams ✋ Hand Detection – enable gesture recognition and interaction 🔥 Fire and Smoke Detection – build early warning systems 🚘 License Plate Detection – power smart parking and traffic analytics 🚗 Car Detection – monitor traffic flow, parking spaces, and more All of these can now be tested instantly using Axelera hardware, straight from your browser.🖥 Try it NowGo to DeGirum AI Hub, select an Axelera-powered model, and start experimenting! Whether you're building a smart city application or an AI-based industrial monitor, AI Hub with Axelera makes prototyping frictionless.

Hi Axelera inferences and compiler available https://hub.degirum.com

I've put together a quick walkthrough of how to set up the Voyager SDK with a Metis M.2 card on the Orange Pi 5 Plus, and wanted to share the full process with you all. This has turned out to be a fantastic low-cost dev setup to start experimenting with edge AI using our Metis platform.Here’s what we cover:Installing the Metis M.2 card Setting up the Orange Pi with the proper OS image Installing the Voyager SDK (using SSH, so you can do it headless) Running a Yolov8L demo with live inferenceEverything is done using standard tools like Balena Etcher, SSH, and a few terminal commands. Whether you’re building a prototype or just exploring Metis on embedded platforms, this setup is a great place to start on a device that's cheap, powerful, and very easy to use.Got questions or ran into snags? Let me know! These resources should help, too:📚 Setup guide reference: https://support.axelera.ai/hc/en-us/articles/27059519168146-Bring-up-Voyager-SDK-in-Orange-Pi-5-Plus📦 Source & SDK: https:/

Just FYI, I’ve been tinkering with the platform design a bit, as I always felt like the code block text input burned your retinas out, being such a big block of white against the dark theme 😆I’ve mimicked GitHub's colour pallette for it, but we can tweak it further if anyone has a preference. (Or if you spot anywhere that’s now messed up because of the CSS changes, let me know),As you were!

I’m using Axelera’s Voyager SDK to run a YOLOv8s object-detection pipeline on video files, and I need to capture every detection into a JSON array for downstream processing. For each detected object, I want to record: frame_index: the zero-based frame number timestamp: the frame’s PTS in seconds bbox: [x1, y1, x2, y2] coordinates class_id: numeric COCO class index class_name: readable label (e.g. “person”) score: detection confidence What’s the cleanest way to hook into the Voyager inference loop, pull out the ObjectDetectionMeta (task name “detections”), convert its boxes, scores, class_ids, and labels into native Python types, attach the FrameResult.src_timestamp, and dump the results into a single JSON file? Example snippets or pointers to the relevant SDK APIs would be greatly appreciated.

Hi Axelera Community,Has anyone here successfully run the Axelera Voyager SDK with ROS 2? I’m currently working on integrating the video_infer node with ROS 2 and would really appreciate hearing from anyone who has tried this setup.My Setup: Hardware: Aetina board SDK Version: voyager-sdk v1.2.5 ROS 2: Humble (on Ubuntu 22.04) Python: 3.10 (using a virtual environment) Model Used: yolov5s-v7-coco Inference Server:python3 inference_server.py yolov5s-v7-coco none --port 50051 --no-display ROS 2 Node:ros2 run axelera_infer video_infer What’s Working: The inference server starts successfully and shows: Server started, listening on 50051 INFO : Using device metis-0:1:0 The ROS2 node launches and shows: [INFO] [video_infer_node]: Starting video inference node... [INFO] [video_infer_node]: Sending inference request... [INFO] [video_infer_node]: StreamInfer started successfully No errors are thrown on either side. Problem: The stream only seems to run a few frames: Stream Playing

Have you seen that ​@Victor Labian’s successfully brought up the Voyager SDK on the Orange Pi 5 Plus with a Metis M.2 module? It took a bit of work—Ubuntu flashing, driver rebuilds, and tweaking the PCIe device tree—but it’s running, and it’s fast. 😃This setup is seriously impressive for something so compact and affordable. Inference runs smoothly, and with some tuning, it could potentially be running advanced models like YOLOv8 without trouble.It’s just an experiment and a test at this point. No big plans around the Orange Pi, but it looks like a really promising start.If you’re experimenting with AI on SBCs—or thinking about it—I’d love to hear what you’re working on. Orange Pi, Raspberry Pi, custom boards, anything edge. 👍 

Hey everyone,Excited to share a brand-new video I put together for anyone getting started with the Metis card and Voyager SDK. Whether you're setting up your eval kit or just curious about how inference works on Metis, this walkthrough is for you.🛠️ What’s inside:Setting up your environment and verifying camera input Running YOLOv5s-v7-COCO on a live camera stream Using download_prebuilt for faster setup Running inference on media files and datasets Benchmarking performance Measuring accuracy with mAP and more📺 You can find this tutorial, and more, on our YouTube ChannelThis guide is based on SDK version 1.2.5 and follows the official Quick Start Guide, but with some extra tips and tricks to get you up and running even faster.💬 Have you tried running inference on Metis yet?Drop your thoughts, questions, or your own setup experiences in the comments, I'd love to hear how it's going for you or help troubleshoot if you're stuck!See you there,Jonas 

We frequently attend trade shows and aim to stand out from the competition. Showcasing compelling demos is a key strategy to achieve this. Which demos would most effectively highlight the unique strengths of our product?

Hi everyone,I am Semih. I am a PhD researcher at University of Tuebingen on Open Source and AI. I am trying to understand Axelera’s hardware and in future I am planning to research legal and regulatory issues related to AI hardware, role of open source hardware, licensing contracts, regulatory compliance especially with Cyber Resilience Act, GDPR and AI Act. I was wondering if anyone is conducting a similar research or are there any pain points or interesting discussions under these themes?Thanks in advance

Hello,I am currently working on lane detection model (YOLOv2) which is in .pt format I have converted to ONNX format.Here are the terminal responses I have got..root@aetina:/voyager-sdk# cd YOLOPv2root@aetina:/voyager-sdk/YOLOPv2# lsLICENSE    data     demo2.py        models            runs   yolopv2-envREADME.md  demo.py  export-onnx.py  requirements.txt  utilsroot@aetina:/voyager-sdk/YOLOPv2# cd modelsroot@aetina:/voyager-sdk/YOLOPv2/models# ls__pycache__  common.py  experimental.py  yolopv2.onnx  yolopv2.ptroot@aetina:/voyager-sdk/YOLOPv2/models# I am having the ONNX format of the model so now I want to deploy the model onto Metis M2. When I am trying to compile the yolopv2.onnx model but i am encountering this issue.root@aetina:/voyager-sdk# /voyager-sdk/bin/voyager-compile models/yolopv2.onnx --output-dir build/yolopv2 --input-shape "1,3,640,640"bash: /voyager-sdk/bin/voyager-compile: No such file or directoryI request to guide me to successfully deploy the model.Note; I am using

This is a quick shoutout out to ​@ChinmayMulgund and ​@oguz for becoming the first members of Axelera AI’s community to earn the rank of Ensign! And practially at the same time (close enough that I can’t tell who got there first).Every activity on the community earns you points, and helps you climb the ranks. The opening rank is a Cadet, which you receive after your first action on the community. But it takes effort to climb the ranks, so a big congratulations to our first two Ensigns for getting there so quickly!Who’s going to be the first Lieutenant, I wonder? 

hi there,i am currently writing up an application with friends for a grant on “ai&art” which is due by tomorrow.we want to do music reactive visuals on edge with multimodal inputs and want to use the Metis for live inference. i think of visuals being an interesting playground since they allow to very freely construct the loads for the chip. if you are interested in collaborating please feel free to reach out! there is no requirement on expertise.  ideas only are also very much welcome.(as well as additional sources for funding or sponsors)hope to have caught your interestbest carlo 

As part of the recent Arduino Day celebrations, our very own ​@Steven Hunsche joined with the Arduino team to demo the Metis PCIe running on a Portenta X8!This is an unmodified LLM, running in real-time, and completely offline.It doesn’t get much more edge than these two device working in combo 😄 And there’s an early access program in the works for anyone who’s interested in getting a closer look at Arduino’s and Axelera AI’s devices in action together: https://axelera.ai/development-systems/arduinoIn the meantime, check out the project on the Arduino Day live stream: 

Wanted to share a quick demo we’ve been playing around with — we’ve got the Metis M.2 up and running on a Raspberry Pi 5, doing real-time vehicle detection using an unmodified YOLOv8 model.Pretty smooth experience, getting them working together. Definitely still early days, so no promises yet on timelines or any kind of future support, but figured it was worth showing that it can be done!Curious — would people here be interested in standalone edge AI running on something like a Pi5? Any particular use cases that come to mind?