search

tensil-ai / tensil

Open source machine learning accelerators
https://www.tensil.ai
Other
348 stars 28 forks source link

Bare-metal deployment #25

Closed nhma20 closed 2 years ago

nhma20 commented 2 years ago

Hi,

How would I go about deploying the tensil accelerator in bare-metal (i.e. without Pynq or Petalinux) on a board like the Ultra96 or PYNQ-Z1? Is it feasible at all? I have been working with FINN (https://github.com/Xilinx/finn) in bare-metal previously and would like to try and compare the two frameworks.

Best regards, Nicolaj

tdb-alcorn commented 2 years ago

We have an embedded driver that can be used in bare-metal deployments. There is an example of how to do it for Ultra96v2 here: https://github.com/tensil-ai/tensil/tree/main/drivers/embedded/ultra96v2

We'll release a tutorial on how to use it in the future, but for now feel free to ping me on our Discord and I can help you get it running.

Discord link: https://discord.gg/TSw34H3PXr

Does that help?

ghost commented 2 years ago

Hi,

We are interested by running bare-metal as well.

Would you happen to have an ETA on the tutorial?

Thanks, Alex

petrohi commented 2 years ago

It is released https://www.tensil.ai/docs/tutorials/speech-robot-part1/

Enjoy!