Testing on x86_64 to port to AARCH_64

TexasInstruments / edgeai-tidl-tools

Edgeai TIDL Tools and Examples - This repository contains Tools and example developed for Deep learning runtime (DLRT) offering provided by TI’s edge AI solutions.

Other

121 stars 27 forks source link

Testing on x86_64 to port to AARCH_64 #35

Closed silver2row closed 1 year ago

silver2row commented 1 year ago

Hello,

Is there a way to build this repo easily? I have tried w/ building it in many, different ways, e.g. make, cmake, w/ build dirs, and w/ docker.

I am not making any headway. The https://github.com/TexasInstruments/edgeai-tidl-tools/tree/master/dockers/dependency_build files here are not working or have been moved.

Seth

P.S. I do not know if people are still interested in these libs. but I have been trying over the course of two months to build them. I have been highly unsuccessful so far. I need hints or ideas at most. Send guidance.

fadilmohd007 commented 1 year ago

Hi,

Sorry to hear that. we appreciate the feedback. For better understanding I would like to know few details on what exactly you are trying to experiment with, I'm unable to deduce following info from the problem statement. 1) Which device are you trying to run the final executable ? 2) Which runtime(python/cpp or both) you are looking forward to ? 3) Where exactly are you stuck ? is it in compilation or execution on target or PC? 4) What is the error message ? 5) Which libs are you referring to ,in the PS section?

silver2row commented 1 year ago

Hello,

BeagleBone AI-64
both
Compilation
All...I am having a complication building

git clone https://github.com/TexasInstruments/edgeai-tidl-tools.git
#export DEVICE=j7
#export DEVICE=am62
cd edgeai-tidl-tools
source ./setup.sh

The above source ./setup.sh command does not build the entire repo. I will get the required error output soon. Sorry I was so vague. I just did not know where to start.

Seth

P.S. The Jupyter Notebooks section builds only so far.

fadilmohd007 commented 1 year ago

Hello,

Currently we are not supporting BeagleBone in edgeai-tidl-tools. But as a work around I can suggest you the following

Compile the model for J721E
Copy the model-artifacts folder generated to the BeagleBone board
Run the execution part

Regarding the setup script:

It is intended only to download necessary python and other packages required for model compilation.
- Run export DEVICE=j7 before sourcing the setup.sh .
For model compilation you can refer https://github.com/TexasInstruments/edgeai-tidl-tools/tree/master/examples/osrt_python#model-compilation-on-pc

silver2row commented 1 year ago

Hello,

Thank you for your time. I will try it again.

Seth