cifar10-ML

Image classification on CIFAR-10 dataset using Machine Learning

INSTRUCTIONS TO RUN THE BEST MODEL:

1. The training binary files and the test binary file should be inside a folder called ‘dataset’.
2. batches.meta.txt should also be inside ‘dataset’ folder.
3. The training binary files should be named as follows: a) data_batch_1.bin b) data_batch_2.bin c) data_batch_3.bin d) data_batch_4.bin e) data_batch_5.bin
4. The test binary file should be names as follows: a. test_batch.bin
5. The .py files should be in the same level as the ‘dataset’ folder.

RUNNING THE PYTHON FILE: Python version 3.6 or 3.7 BEST MODEL: MODEL 1 HOG + SVM CLASSIFIER Filename : model_1_hogSvm.py It should be run from Anaconda Prompt with an argument (‘preTrain’ or ‘forceTrain’). Example:

ForceTrain – Starts Training from scratch PreTrain – Loads previously trained classifier from folder if found and uses it to predict the test data.

2nd BEST MODEL : MODEL 2 VGG + RFC CLASSIFIER Filename : model_2_vggRFC.py Similar instructions as model 1

OUTPUT: The testing Accuracy is displayed on the console. The predictions are saved as a .csv file in the folder in respective model names.

NOTE: If the pre-trained files are deleted from the folder, ‘ForceTrain’ is enforced automatically.