modified the loss function for _cnn.py

[Vasudeva-bit]

Reference Issues/PRs

Not an issue fix

What does this implement/fix? Explain your changes.

I have modified the loss function of the _cnn.py file. The current loss function for the _cnn.py file which is used to classify between two classes is RMSE but the appropriate loss function for training a model for binary classification is binary_crossentrophy. This loss function can significantly improve the training and performance of the estimator.

Does your contribution introduce a new dependency? If yes, which one?

No

Any other comments?

No