Instruction to use the code package.

Requirements:
1. Python. The code was tested in versions: 3.7.6 & 3.7.4
2. PyTorch. The code was tested in versions: 1.3.1 & 1.2.0
3. Numpy. The code was tested in versions: 1.17.2 & 1.19.2
4. ipython. The code was tested in versions: 7.8.0 & 7.19.0
5. sklearn. The code was tested in versions: 0.21.3 & 0.22.1


Notes: 
1. The Indian Pine dataset is given with the package.
2. Two scripts need to be run.
3. The script testSCE_IP.py will create a ranked feature set using the training data.
   The feature set with top 80 features will be written in "./FeatureSet/IndianPine/IndianPine_features.p"
   which is a pickle object.
4. The script "SupervisedCentroidencodeClassifierPyTorch_IndianPine.py" will do the classification
   using the top n features where n=1,2,3,4,5,10,20,40,60,80

How to run the code:
1. Download the code.
2. Make sure all the requirements are satisfied.
3. To run the script from ipython use the commands:
    a. ipython
	b. run testSCE_IP.py (This will create a ranked feature set using the training data. This will take time.)
	c. run SupervisedCentroidencodeClassifierPyTorch_IndianPine.py (This script will run classification)

