Skip to main content

Metis M.2

Maximise your Metis M.2 – connect, ask, and share to optimise your AI projects with this powerful edge accelerator

  • 25 Topics
  • 219 Replies
25 Topics
M
Cadet
mgrunerCadet
asked in Metis M.2
Low performance of the Metis M2 on the Raspberry Pi5

Hi! I'm testing the Metis M2 on a Rapsberry Pi 5 using the Ubuntu 22.04 container. I can successfully get results using the inference test suggested by the article:./inference.py yolov8s-coco-onnx ./media/traffic1_480p.mp4 Unfortunately, I get around 4FPS. Is this expected? If I profile the device and the host separately (using --show-host-fps and --show-device-fps) I see that the device is running at ~800FPS while the host is the one bottlenecking at ~4FPS.I tried enabling GLES processing using export AXELERA_OPENGL_BACKEND=gles,3,1but, unfortunately, it makes not much of a difference. I also tried setting the PCI to gen3 by setting the following in the /boot/firmware/config.txt, without any luck:dtparam=pciex1_gen=3As a reference: these are the instructions I'm following:- https://support.axelera.ai/hc/en-us/articles/26362016484114-Bring-up-Voyager-SDK-in-Raspberry-Pi-5- https://support.axelera.ai/hc/en-us/articles/25953148201362-Install-Voyager-SDK-in-a-Docker-ContainerIs this perfo

433
M
Cadet
A
Cadet
AfikCadet
asked in Metis M.2
Yocto Driver\Layer for Metis

Hi, We are considering adding an AI Accelerator for our board (i.MX8, yocto).Do you support yocto OS?Is there any reference design or example that runs Metis on yocto platform? Thanks,Afik

51
Steven Hunsche
Axelera Team
K
Cadet
ku15921Cadet
asked in Metis M.2
Error installing metis-dkms

Hi,I'm having trouble installing the metis-dkms driver. Can you help me?The device is a NanoPC-T6 from Friendly Elec, and the operating system is Ubuntu 22.04, also provided by Friendly Elec.I'm leaving the error message and OS information below. If there is any more information I should provide, please let me know. Thank you. The Error:$ sudo dpkg -i metis-dkms_0.07.16_all.deb(Reading database ... 155441 files and directories currently installed.)Preparing to unpack metis-dkms_0.07.16_all.deb ...Deleting module metis-0.07.16 completely from the DKMS tree.Unpacking metis-dkms (0.07.16) over (0.07.16) ...Setting up metis-dkms (0.07.16) ...Loading new metis-0.07.16 DKMS files...Building for 6.1.99Building for architecture aarch64Building initial module for 6.1.99ERROR (dkms apport): kernel package linux-headers-6.1.99 is not supportedError! Bad return status for module build on kernel: 6.1.99 (aarch64)Consult /var/lib/dkms/metis/0.07.16/build/make.log for more information.dpkg: error pro

20112
K
Cadet
dbrenes
Cadet
dbrenesCadet
asked in Metis M.2
Metis M.2 Not Recognized by lspci on Intel system

Hello! I’ve been trying to connect and use the Axelera Metis M.2 card with no success. The same M.2 on a different PC was recognized correctly by using:lspci | grep AxeleraBut now the card cannot be recognized at all on the new system. Here are some specs of the current host:Dell 5820 CPU: Intel(R) Core(TM) i7-7820X CPU @ 3.60GHz RAM: 32 GB GPU: 2 Quadro RTX 5000 OS: Ubuntu 24.04 I’ve also tried lspci | grep 1f9d with no success. The Metis driver is already installed and loaded with modprobe.I'd appreciate any help or insight on additional steps I could try.Additionally, this is the adapter I’m using for the Axelera card: 

264
dbrenes
Cadet
G
Cadet
GiodstCadet
asked in Metis M.2
WASB model for tennis ball detection

How can I deploy WASB model on metis m2?

222
G
Cadet
Steven Hunsche
Axelera Team
Steven HunscheAxelera Team
posted in Metis M.2
4k fruit detection using Orange Pi

Have you seen our fruit demo? If not, ​@David explains it amazingly here.Basically, it is running 3 different yolov8 models on a single chip to process 4k camera input to segment fruit, while only using 640x640 models!I love this demo and use it to evaluate any system that comes across my desk! So, obviously, when ​@Victor Labian posted the setup for the Orange Pi here, I just had to try it out! And it works perfectly straight after following the instructions!I shaded the fruit to look different from the original to show the accuracy of the segmentation.If you want to run this demo for yourself, you can! It’s available in the examples folder of your SDK, called fruit_demo.py. Let me know what you think! Or, as ​@Spanner suggestedA Fruitful and A-peel-ing Orange Pi Experiment that Didn't Go Pear-ShapedHere's what's current on the Axelera grapevine - a core functionality of the Orange Pi with Metis is detecting fruit in real time! Super smoothie AI pipe-lime with bananas and apples being

461
Spanner
Axelera Team
M
Cadet
ma926Cadet
asked in Metis M.2
Metis M.2 1.3.0 firmware upgrade

Hi! I wanted to upgrade my M.2 card with latest firmware 1.3.0 but when trying to invoke command from instruction https://github.com/axelera-ai-hub/voyager-sdk/blob/release/v1.3/docs/tutorials/firmware_flash_update.md#flash-the-board-controller-firmware, I got an error:Do you know how this can be resolved? I’m using Intel i5 processor.

371
Spanner
Axelera Team
D
Cadet
doggenshozenCadet
asked in Metis M.2
Can I use my GPU memory with Metis M.2 for local LLM?

This may not be possible because of how the software and hardware work together. I’m new here, and not traditionally a SW engineer, but I have a question/ suggestion. I noticed when I attempted to run the example local LLM inference scripts in the SDK that I can’t run them with only 1GB of DDR memory on the card. I think the example needs to be edited so that the model can be run on 1 GB of DDR (a model like tinyllama or similar that’ll actually fit in the 1GB of space.)After thinking about it, I do have an intel Raptor Lake integrated GPU, and an external NVIDIA A500 from Adlink Technologies. Is there any way to use the memory in the integrated CPU or the thunderbolt connected A500 GPU to run the LLM inference demos by using the DDR memory in the other devices? That would be very very useful!

150
D
Cadet
S
Ensign
saravananEnsign
asked in Metis M.2
Blank screen issue with Metis M.2 Eval System with Aetina RK3588 Industrial Motherboard

i am using Metis M.2 Eval System with Aetina RK3588 Industrial Motherboard . i connected everything to it keyboard, monitor and mouse and power supply . the monitor is showing blank screen . but yesterday it is working . why it is not working now . is there any solution. 

31629
S
Cadet
S
Cadet
sifat0433Cadet
asked in Metis M.2
Input/Output Error Running Inference with Metis M.2 on ASRock Z170 Pro4S

Hello everyone,I’m new to working with this hardware and recently installed the Metis M.2 chip (Firmware version: v1.2.0-rc2+bl1-stage0) on an ASRock Z170 Pro4S motherboard. The device was detected successfully.I’ve installed the Voyager SDK and attempted to run the following command from the Quick Start Guide:./inference.py yolov5s-v7-coco dataset --no-display However, I encountered an input/output error when executing the command. I’ve included the full error message below for reference.I would appreciate any guidance or suggestions you may have to help resolve this issue.Thank you in advance for your support. INFO    : Using default val dataset INFO    : Using device metis-0:5:0 INFO    : Network type: NetworkType.SINGLE_MODEL INFO    : Input INFO    :   └─detections INFO    : Imported DataAdapter ObjDataAdaptor from /home/tud/voyager-sdk/ax_datasets/objdataadapter.py Stream Playing:  71%|██████████████████████▏        | 5/7 [00:00<00:00,  6.26/s][AxeleraDmaBuf.cpp:234] UIO_IOCTL

11510
Spanner
Axelera Team
I
Cadet
IAKCadet
asked in Metis M.2
Problems with network inference. Uio wait kernel failed with return code -1406.

Hello!I am facing some problems tring to run a network inference on the Metis M.2 with Raspberry Pi via Docker. To try the inference I have followed the steps in the guide (https://support.axelera.ai/hc/en-us/articles/26362016484114-Bring-up-Voyager-SDK-in-Raspberry-Pi-5) but after trying to run the inference.py, I face the following errors:  (venv) root@iak:/home/voyager-sdk# ./inference.py yolov8s-coco-onnx ./media/traffic1_480p.mp4INFO    : Using device metis-1:1:0WARNING : Failed to get OpenCL platforms : clGetPlatformIDs failed: PLATFORM_NOT_FOUND_KHRWARNING : Please check the documentation for installation instructionsINFO    : Default OpenGL backend gl,3,3 overridden, using gles,3,1INFO    : Network type: NetworkType.SINGLE_MODELINFO    : InputINFO    :   └─detectionsStream Paused:  57%|██████████████████████████████████████████████████████▊                                         | 4/7 [00:00<00:00,  5.56/s][ERROR][axeWaitForCommandList]: Uio wait kernel failed with return c

15120
Spanner
Axelera Team
C
Ensign
ChinmayMulgundEnsign
asked in Metis M.2
Guidance on Running Lane Detection Model (culane_res18.onnx) on Metis M2

Hello,I am currently working with a lane detection model named culane_res18.onnx, and I would like to know how I can run this model on the Metis M2 device.Here are some details about the setup:The model is culane_res18.onnx, which is a lane detection model. The training dataset is from the CULane dataset, and I have the link to the dataset here. The model file can be found here.Could you kindly guide me on how to run this lane detection model on Metis M2? I would appreciate any step-by-step instructions or basic guidance on how to deploy this model, considering that I have both the ONNX model and the dataset.Looking forward to your response!Best regards,Chinmay Mulgund

35224
H
Axelera Team
S
Ensign
saraEnsign
posted in Metis M.2
Ubuntu 24.04 support for RPi5, no container

Hello!!Do you have any news about support for Ubuntu 24.04? I wanted to use axelera on raspberry pi cm5 however there is no support for 22.04 and I wanted to run it on the host, not in a docker container.Thank you 😊

301
Spanner
Axelera Team
C
Ensign
ChinmayMulgundEnsign
asked in Metis M.2
AXELERA M2 Not Detected – “No target device found in lspci output”

Hello,I’m having trouble getting my AXELERA M2 to be recognized under Linux. Until recently everything was working fine; now when I try to run an inference the SDK can’t find the board. I’ve already tried rebooting, but the issue persists.aravind@aravind-H610M-H-V2:~/Desktop/voyager-sdk$ source venv/bin/activate(venv) aravind@aravind-H610M-H-V2:~/Desktop/voyager-sdk$ ./inference.py yolov5s-v7-coco chinmay/Vehicle-CV-ADAS/assets/challenge_video.mp4 --save-output ~/Desktop/voyager-sdk/chinmay/saved_videos/yolov5s-v7-coco.mp4ERROR : No target device found in lspci outputERROR : AXR_ERROR_CONNECTION_ERROR: No target device found in lspci output 

693
Spanner
Axelera Team
P
Cadet
pautu227Cadet
asked in Metis M.2
Trying to setup rpi5 with metis m2

​@Victor Labian I’m also trying to setup rpi5 with metis m2. I have question what host OS you was using and have you changed anything in config.txt?

30629
Spanner
Axelera Team
S
Cadet
SaitejaCadet
asked in Metis M.2
Deploying the YOLOPv2 model on to Metis M2 for lane detection

Hi, I am also working on same thing deploying the YOLOPv2 model on to Metis M2. I have downloadded the data set from bdd100k and arranged in this manner.voyager-sdk/├── data/│   └── yolopv2_dataset/│       ├── images/│       ├── labels/│       ├── cal.txt│       ├── val.txt│       └── data.yaml├── customers/│   └── my_yolopv2/│       └── yolopv2.pt└── yolopv2-custom.yamlso when i am trying to deploy the model using yaml I am getting this error(venv) aravind@aravind-H610M-H-V2:~/Desktop/voyager-sdk$ ./deploy.py customers/my_yolopv2/yolopv2.yamlINFO    : Using device metis-0:1:0INFO    : Detected Metis type as pcieINFO    : Compiling network yolopv2-custom /home/aravind/Desktop/voyager-sdk/customers/my_yolopv2/yolopv2.yamlINFO    : Compile model: yolopv2-customINFO    : Imported DataAdapter ObjDataAdaptor from /home/aravind/Desktop/voyager-sdk/ax_datasets/objdataadapter.py/home/aravind/.cache/axelera/venvs/93f45ae3/lib/python3.10/site-packages/torch/serialization.py:779: UserWarning: 'to

946
B
Axelera Team
F
Cadet
floritCadet
asked in Metis M.2
Metis M.2 on Raspberry Pi 5 - Driver Installation and Docker Setup Issues

hello I hijack this a little bit … once I type “lspci” it gave me following output on a pi 5 0001:00:00.0 PCI bridge: Broadcom Inc. and subsidiaries BCM2712 PCIe Bridge (rev 30)0001:01:00.0 Processing accelerators: Axelera AI Metis AIPU (rev 02)0002:00:00.0 PCI bridge: Broadcom Inc. and subsidiaries BCM2712 PCIe Bridge (rev 30)0002:01:00.0 Ethernet controller: Raspberry Pi Ltd RP1 PCIe 2.0 South Bridge how to activate the device and getting to work with my python3 script  

22631
F
Cadet
M
Cadet
MehmetYigitCadet
asked in Metis M.2
Metis M.2 Accelerator Can't Recognized by Jetson AGX Orin

Hello.I recently received the m.2 metis accelerators and did a test on the boards provided by Axelera and got results close to the published benchmarks. It was succesful.I then connected these accelerators to the Orin (via the m.2 slot) in order to use them also for the my Jetson AGX Orin environment. But in the voyager-sdk installation i got “WARNING: Failed to refresh pcie and firmware” error. The installation completes but the accelerator device does not show up on my system. In lspci -tv command I get the output attached to this post. Can’t see metis there.I didn't see any information about the compatibility of these accelerators with Jetson systems (my m2 slot is M-key). If they are compatible, can you help me to fix the installation?Thanks :)

1308
Spanner
Axelera Team
D
Cadet
DrouenCadet
asked in Metis M.2
Troubleshooting installation

Hello my name is Kevin, i am trying to work on a project that the m.2 is going to interfere with my cpu and later on maybe my gpu.is there a way, and another question whats the command to go on rtsp?  my system is:AMD 9950X3D CPUAMD 7900XTX GPU48 gb DDR5 on 8200MHZMetis Axelera AI on an m.2 port. 

8412
Dr Data
Ensign
M
Cadet
MartianCadet
posted in Metis M.2
Testing and M.2 card on an AMD system, not recognized by lspci

This is a very interesting thread as I’m testing and M.2 card on an AMD system as well and am having what seems to be the exact same problems. Things added / tried: kernel command line includes md_iommu=off pcie_aspm=off update-pciids → resulted in card having the correct name in lspci : Axelera AI Metis AIPU (rev 02) axdevice --refresh -v results in: INFO:axelera.runtime.axdevice:Removing 0000:20:04.0INFO:axelera.runtime.axdevice:PCIE rescan0000:25:00.0 : Axelera AI Metis AIPU (rev 02)INFO:axelera.runtime:Found PCI device: 25:00.0 Processing accelerators: Axelera AI Metis AIPU (rev 02)INFO:axelera.runtime:Found AIPU driver: metis                  90112  0WARNING:axelera.runtime:4PCI device count mismatch: lspci=1, triton=0(removed the traceback) just tried rerunning install.sh which mentions at the end: building operators refreshing pcie and firmware WARNING: Failed to refresh pcie and firmware  tried removing metis-dkms / rmmod metis and reinstalling results in same warning as ab

824
D
Cadet
O
Ensign
oguzEnsign
asked in Metis M.2
Transforming the Attention mechanism into Convolutional

Hello,I know that axelera does not work in the attention mechanism. For example, can we convert the attention mechanism in the yolov11 architecture to convolution? Have you done any research or read on this subject? Does Softmax structure work in Axelera? (as far as I know, no) Is there a way to turn Softmax into a working structure?

682
B
Axelera Team
O
Ensign
oguzEnsign
asked in Metis M.2
How does Axelera speed up models?

Hello, is there a detailed document where I can learn about the Axelera chip architecture and how to do the quantization in it?Thanks

522
B
Axelera Team
O
Ensign
oguzEnsign
asked in Metis M.2
Tested at Yolov8x and Yolo11

Hello,I wanted to test the Yolo11n model. But I couldn't do this because I got an error in the calibrating part. Is this because of the use of attention mechanism in Yolo11 and Yolo12?Error: ValueError: Some modules couldn’t be simplified: <… activation.Softmax> <…. matmul.MatMul>In addition, when I looked at the ready yaml files on voyager sdk github, I saw that the yolov8x model was not available. Is there a special reason for this? Because when I prepared and tried a new yolov8x yaml file, I got the similar error again. But as far as I know, the attention mechanism is not used in Yolov8x.Thanks

19310
Berrisius
Cadet
gurumeditation
Ensign
gurumeditationEnsign
asked in Metis M.2
Considering the M.2 module – what can it realistically handle?

I’m scoping out the Metis M.2 module for a compact edge AI project and wondering how capable it really is. I’m mainly interested in real-time object detection on 1080p video streams, maybe 2–4 cameras max.Can it handle multi-stream inputs in real-time, or is that better suited to the PCIe version?

1134
Spanner
Axelera Team
Spanner
Axelera Team
SpannerAxelera Team
posted in Metis M.2
First tasks when getting started with a Metis M.2

It’s actually not that complicated getting started with an M.2 module, and there’s a good chance that the Metis isn’t the first M.2 you’ve ever had in your hands.But it’s still good policy to familiarise yourself with both the hardware and the software when you’re ready to unbox that little AI beauty.To help with that, we’ve got a series of articles to walk you through…Installation: Handling and installing the M.2 in your host system, and a great tip on how to confirm there’s suitable power getting to it once it’s in place Integration: Top tips on figuring out whether the host PC (or system or platform) is nicely compatible with the needs of the Metis accelerator. Voyager SDK: Once it’s installed and ready, all you need is the Voyager SDK.This is all just to get you started, and then the real work begins! Make sure you share what you’re building right here on the Axelera community 👍

490
Spanner
Axelera Team

Badge Winners

Show all badges
Code of Conduct

Address

HTC5, High Tech Campus
5656 AE Eindhoven
The Netherlands
Email: info@axelera.ai

Menu

AI Accelerators

AI Software

Metis Evaluation Systems

Partners

Product Inquiry

Company

Our Story

Our Team

Careers

News

Contact

Reducing CO₂ with
Axelera’s Forest

Follow Us

LinkedIn

Twitter

Sign Up for Our Newsletter

EU flag Axelera AI is co-funded by the European Union’s Horizon research and innovation programme.

Privacy Policy  |  Cookie Policy  |  Terms of Sale  |  Vulnerability Disclosure Policy

Copyright 2025 © All rights Reserved.