#GraTO ####file structure ###draft version
├── architect.py // architecture parameter calculation and optimization
├── args.py
├── final_operation_test // opeartions used in GraTO
├── genotype.py // select operations you want to use in the searching process
├── layer.py // pairnorm code
├── LDA_loss.py // just ignore it
├── Mad_measure.py // calculation of MAD value
├── model.py // basic block code without architecture parameter
├── new_measure.py // metrics
├── new_model_search.py // basic block code with architecture parameter
├── new_train_newmeasure.py // train derived structure code
├── new_train_search_newmeasure.py // structure searching code
├── normalization.py //ignore
├── start.py // here set all parameters and train your derived model
├── utils.py
If you have any questions, please email me so I can get to know.