#####################################################################################
# Classification Experiment with 1024 Feature Dimensions on the Penultmate layer
#####################################################################################

# OPL
##########

1. To run the Classification Experiment:
	- Execute OPL_Base_1024D.py to save the model as OPL_BASE_1024D_model.h5
2. Generate OPL features:
	- Run getOPL_Features.py to extract the 1024 Dimensional Features from the Penultimate layer for the CIFAR-100 Training and Test Data
3. Compute Average class dot products and Intra-Class Compactness:
	- Use ComputeDotProduct.py to calculate the Average class dot product scores
	- Use computeCompactness.py to compute Intra-Class Compactness
	- Ensure to update the path to the location where the Training Features and Labels are saved


# OPL+ICR or # OPL+DCR
##############################
1. Generate OPL+ICR or OPL+DCR Features:
	- Run getOPL+ICR_Features.py or getOPL+DCR_Features.py respectively to extract the 1024 Dimensional Features from the Penultimate layer for the CIFAR-100 Training and Test Data
2. Compute Average class dot products and Intra-Class Compactness:
	- Use ComputeDotProduct.py to calculate the Average class dot product scores
	- Use computeCompactness.py to compute Intra-Class Compactness
	- Ensure to update the path to the location where the Training Features and Labels are saved
3. To run the Classification Experiment:
	Execute OPL+ICR_DCR.py and update the path to the location where the OPL+ICR or OPL+DCR Features and Labels are saved


#####################################################################################
# OOD Experiment with 1024 Feature Dimensions
#####################################################################################

# CIDER
##########
 
1. Run OPL_Base_1024D.py first to save the model CIDER_BASE_1024D_model.h5 for feature extraction
2. Generate CIDER CIFAR-100 in-distribution features:
	- Execute getCIDER_Features.py
3. Generate out-of-distribution features:
	- Download the OOD data from https://github.com/deeplearning-wisc/cider
	- Run getOOD_Features.py to generate the OOD features
4.To run the OOD Experiment:
	- Execute OOD_Exp.py
