Black screen for models with resolution > 100 [STM32F746G-DISCO]

mit-han-lab / tinyengine

[NeurIPS 2020] MCUNet: Tiny Deep Learning on IoT Devices; [NeurIPS 2021] MCUNetV2: Memory-Efficient Patch-based Inference for Tiny Deep Learning; [NeurIPS 2022] MCUNetV3: On-Device Training Under 256KB Memory

https://mcunet.mit.edu

MIT License

792 stars 130 forks source link

Black screen for models with resolution > 100 [STM32F746G-DISCO] #103

Open rahulvigneswaran opened 5 months ago

rahulvigneswaran commented 5 months ago

For models with original resolution > 100, we encounter a black screen instead of the prediction result.

This could be happening because the loadRGB565LCD function in the main.cpp is trying to fit all these pixels in the display.
We tried to reduce the resize factor in the loadRGB565LCD function but the issue persists.
A simple way to get it to work is to use half the resolution but it doesn't look like the correct way to do it.