Update README.md

README.md

@@ -6,3 +6,14 @@ In our country there is a huge crop wastage due to not been able to detect disea
 ## MOTIVATION OF THIS PROJECT:
 Generally while applying CNN models problem is that number of parameters required are huge and it takes a lot to time to predict it with an application.Some models such as ALexNet takes 238M params,VGG-16 has 654M params.
 So to cure the problem instead of normal convolution we have used depthwise separable convolution which reduces number of params a lot. 
+
+## Dataset used
+https://www.kaggle.com/datasets/alyeko/potato-tomato-dataset
+
+## Models used
+1) General Convolution + Hidden layers
+2) Xception Model  
+3) MobilenetV2
+4) Inception V3 
+5) Modified Inception V3
+