this dataset contains natural images of size 150*150 for 6 categories. the training set contains 6 folders around 2000 for each category, the testing set contains 3000 images. the goal is to build a model that can recognize the category of each image in the testing set