I want to create a gender and age detection algorithm for maixduino board . I have been creating custom models using axelerate framework . i have used different architectures like resnet50, densenet ,mobilenet architectures but none of them is giving any good models.
Is there any other way to train models for maixduino?
what should be ideal range of size for a model?
Can you please help me . I have been stuck in this for many days