author Stephanie
All implementation based on corresponding model paper. It is suitable as a reference for reading papers.
To modify options such as is_gpu and is_resume or super parameters such as LR, please modify config.py`
Some strategy changes need to modify the source code of train.py (e.g. modify the learning rate schedule)
PS:ShuffleNet accuracy is not good, welcome to practice model tuning.