In this competition you will be predicting the type of defects found in alloys. Images are named with a unique ID. You must classify the defects in the test set.

Each image may have no defects (not included in the train.csv), a defect of a single class, or defects of multiple classes. For each image you must segment defects of each class (Class = [0, 1, 2, 3, 4]).

The segment annotations are also provided for training purposes.

 the data set will be provided through a drive link 
The file uploaded here for publishing sake