how to use mobilenetV3 small to train my custom dataset for object detection?

1.3k views Asked by At

i want to train my dataset using mobilenetv3 small for object detection using google Colab. but its not provided in the model zoo. and i cant find the config file to train the model. is there any other way ? a link to the config file will help. I tried using the provided google Colab link in TensorFlow repository but I couldn't train my dataset on it.

1

There are 1 answers

0
jansary On

mobilenetv3 is a "backbone" architecture and is used for feature extracting. if you want to use it for object detection, you have to use it with object detection algorithms like ssd, faster_rcnn and etc. example: https://colab.research.google.com/drive/1eAUjzV3nZXkUXi0spPg6zUHtJWaUSzFk?usp=sharing