Inicio Explorar Axuda
Iniciar sesión
RiceChen0
/
r-tinymaix
réplica de https://github-proxy.rt-thread.io/RiceChen0/r-tinymaix.git
2
0
Fork 0
Ficheiros Incidencias 0 Wiki
Rama: main
Ramas Etiquetas
r-tinymaix
/
examples
/
maixhub_detection_fomo
RiceChen 2b0f91e159 init tinymaix code %!s(int64=2) %!d(string=hai) anos
..
include 2b0f91e159 init tinymaix code %!s(int64=2) %!d(string=hai) anos
CMakeLists.txt 2b0f91e159 init tinymaix code %!s(int64=2) %!d(string=hai) anos
README.md 2b0f91e159 init tinymaix code %!s(int64=2) %!d(string=hai) anos
fomo.c 2b0f91e159 init tinymaix code %!s(int64=2) %!d(string=hai) anos
main.c 2b0f91e159 init tinymaix code %!s(int64=2) %!d(string=hai) anos

README.md

MaixHub FOMO object detect demo

  • Train your model on MaixHub first, select tinymaix plaform and use fomo detection method when create training task
  • Download model file and unzip, copy model-*-final.h to this directory, rename to model_final.h, and copy model_info.h too
  • Change config in main.c according to your model(params can be found in report.json)
  • Use python3 ../../tools/img2c.py test_img.jpg width height to generate test input image c file and copy to this directory
  • mkdir build && cd build && cmake .. && make, then execute ./fomo if you use Linux

Change include/tm_port.h configs like TM_MAX_KCSIZE according to your model size.