stacked-auto-encoders-recommendor-system-
the data preprocessing segment is same to that of the boltzman machine code i have in my repository..the google lens dataset is used (100k one). the auto encoder has been build up from the scratch and got quite a good loss.. if u need explanation of it in brief ,contact me in @hack.it29 thanks to the blog(https://blog.keras.io/building-autoencoders-in-keras.html) it helped me alot on gettin the intuition behind the scenes.