Didnt understand the need of conversion to float

[kushal1998 (kushal agrawal)]

here why we are not vectorizing ytest and ytrain?why we are converting them to float…whereas xtest and xtrain are in intigers?

[rachitbansal2500 (Rachit Bansal)]

why we are converting them to float

See, our model predicts a float value representing the ‘confidence’ (just like in Logistic Regression) when we pass an x_data into it. To compute loss we would have to take the difference between the actual values (which are stored in the y_train) and the values predicted by our model, thus both of them must be in the same data type, i.e., float.

here why we are not vectorizing ytest and ytrain?

It’s a case of binary classification, we do not need to vectorise them because our network is predicting one single value also which is being compared to the actual value.