Hello,
I’ve been trying to install and use numpy, opencv-python and tensorflow on the AXC F 2152 controller and I have not been very successful. Is there a way to work around this?
I’m not sure if this helps, but there are instructions for installing similar python libraries on the AXC F 3152 in this tutorial:
How to set up AXC F XT ML 1000 After you have added the files libedgetpu.so.1 and libflatbuffers.so.1 to the \\usr\\lib directory, check whether the AXC F XT ML 1000 module is accessible: The tutorial relates to the ML 1000 module, which cannot be used with the AXC F 2152, but perhaps some of the instructions for installing python modules might help in your case.
If not, can you describe the problem(s) you are seeing?
Hello,
is this tutorial not available anymore? I get a 404 error when accessing that page.
regards,
Tom
Hi Tom, unfortunately a recent re-design of the Info Center has broken most of the links to the old site.
Here is that tutorial on the new site:
How to set up AXC F XT ML 1000 After you have added the files libedgetpu.so.1 and libflatbuffers.so.1 to the /usr/lib directory, check whether the AXC F XT ML 1000 module is accessible:
Hi Martin,
thanks, one more question:
The PLCnext Store Python Basic package says “This app works up to the firmware version 22.6”.
Our 3152 has 23.0.2 installed already.
Am I running into issues if I use that python package from the store? Or should I install python manually into a balena container?
Which phython version is recommended for the TX ML 100 examples?
regards,
Tom
Hi Martin,
meanwhile I’ve found the relevant passage in the document that answers my question:
“Note: For controllers with firmware versions from 2022.6, use a container (e. g., docker) including Python 3.9.”
Since firmware 23.0.2 has python 10 installed, I guess I’ll need a container setup with python 3.9 to run the ML examples.
regards,
Tom