Hello @zauberflote1! Thanks for sharing.
Made any progress on this package?

@Alex-Kushleyev Correct, I just used the last 4 more powerful cpus. That was enough of a performance boost to run my optimization at the frequency I was hoping for. I didn't try using all 8 threads, I can give it a shot if that's useful information for you

@tahawaru Okay, that shows you have 1.14.0-2.0.78. If you want to build 1.14.0-2.0.63 then you need to check that version out instead.

@modaltb

Hey Travis,

Thank you very much for your explanation and for passing on the link.
I understand how it was done now.

I will consider what is the next possible option for the project.  
I appreciate your time.

Best Regards,
Leon

@nickk14

Awesome, glad to hear it! Yeah, we've actually mounted fans to our boards that run multiple models simultaneously because the thermals definitely get up there. Might be worth thinking about depending on your application.

All the best, let me know if I can help out in any other way.

Thomas

@Moderator Thank you for the response.

I did come across this product of yours here: https://www.modalai.com/products/mdk-m0135?_pos=1&_sid=e4d19a51b&_ss=r that should work great after looking through some of your user guides and technical documentation.

@eric

Sorry for the late response on this!!

Your best bet would be to just try it out, I can't say without trying it myself. It's possible it works, it's possible something fails. I've never had to deal with this so I can't say for certain.

It's also worth pointing out that voxl-tflite-server is more meant to be an example of how TFLite models can be run on VOXL and not an in-depth ML framework for deploying custom models. If voxl-tflite-server doesn't support it, by all means fork the repository and build out the functionality to be able to. Even better, make a pull request with that functionality and I'll test and approve it!

Hope this helps,

Thomas Patton

@david-devassy

Hey David,

We do support deep learning inference on VOXL2! You can check out this page on our docs to read more about our deep learning capabilities. Out of the box our YoloV5 model gets around 30 frames per second. You also mention object avoidance and SLAM using path prediction. We have our own implementation of Visual Object Avoidance (VOA) which you can read about here.

You could certainly create custom implementations of both of these on VOXL! However I want to highlight that our code for both of the above is C/C++ and not in Python for speed reasons. I can't give you a reasonable estimate about Python performance but it would likely be slower.

Hope this helps! Let me know if you have any other questions.

Thomas Patton
thomas.patton@modalai.com

@dsupplee yes, the images on VOXL 2 are unsigned. please see the VOXL 2 kernel build guide here: : https://docs.modalai.com/voxl2-kernel-build-guide/

@Moderator Got it. I have mailed already to contact@modalai.com . Please expedite my request and let me know.

@brinkap unfortunately, there is no CAN capability on VOXL 2. You would need to have an adapter PCB that supports CAN connects to one of the available interfaces, and then write a Linux driver that enumerates as can0

@ErezI

I think we did it, but i'll try again.
Thanks

@camharrisvayu At this point it may be worth submitting an RMA for this flight deck and we can take a look: https://www.modalai.com/pages/rma

@camharrisvayu Ok, we should be able to close this issue -- I flashed 1.0.0 and uploaded params for the sentinel model and that made it so the RC inputs could control the drone -- not sure which parameter made the difference, but I may investigate and report back if I find out.

@brycek did you eventually figure out? I need to create a full image as well