TensorFlow Lite Object Recognition on the Raspberry Pi 4, ca

General project help for Adafruit customers

Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.
Locked
User avatar
SpencerLhasa
 
Posts: 5
Joined: Mon Sep 13, 2021 3:48 am

TensorFlow Lite Object Recognition on the Raspberry Pi 4, ca

Post by SpencerLhasa »

For the TensorFlow Lite Object Recognition on the Raspberry Pi 4, can you use a hdmi display instead of the Adafruit BrainCraft HAT with the 240x240 display?

User avatar
mikeysklar
 
Posts: 13936
Joined: Mon Aug 01, 2016 8:10 pm

Re: TensorFlow Lite Object Recognition on the Raspberry Pi 4

Post by mikeysklar »

Are you planning on using your BrainCraft HAT for vision projects? It looks like FBCP (framebuffer copy) mode is required in that case. You should still be able to use HDMI, but it will be at low resolution (240x240).

If you are not doing vision projects you can just run console mode.

Code: Select all

sudo python3 adafruit-pitft.py --display=st7789_240x240 --rotation=0 --install-type=fbcp

Code: Select all

sudo python3 adafruit-pitft.py --display=st7789_240x240 --rotation=0 --install-type=console

User avatar
SpencerLhasa
 
Posts: 5
Joined: Mon Sep 13, 2021 3:48 am

Re: TensorFlow Lite Object Recognition on the Raspberry Pi 4

Post by SpencerLhasa »

Thank you. While waiting for the BrainHat, I just tried running the program with using just the HDMI.

Now that I have the BrainHat and the TFT configured:
" The advanced way is to install a kernel module to add support for the TFT display that will make the console appear on the display."
https://learn.adafruit.com/adafruit-bra ... le-install

If I VNC from my laptop over to the Pi, my display on the laptop matches the 240x240 resolution (like you said) and I can start and stop the program using VNC.

If I remote desktop from my laptop to PI, I can get a "full resolution" display on my laptop
the TensorFlow program using the remote desktop cannot access the TFT and crashes.

If I switch the configuration of the TFT to be installed:
The easy way is to use 'pure Python 3' and Pillow library to draw to the display from within Python.
will I be able to run the TensorFLow program using the remote desktop?

The reason for this request/change, is that I am trying to learn python by running the TensorFlow program under the PyCharm IDE, hence a large resolution screen (under Remote Desktop, not VNC to the 240x240 screen) is necessary.

Thanks

User avatar
mikeysklar
 
Posts: 13936
Joined: Mon Aug 01, 2016 8:10 pm

Re: TensorFlow Lite Object Recognition on the Raspberry Pi 4

Post by mikeysklar »

Based on how you want to work I think it is worth trying the "pure Python 3" easy way and see if it is workable. I actually do not know for sure, but I think it is worth trying.

Locked
Please be positive and constructive with your questions and comments.

Return to “General Project help”