Movie recommendation
Try it right now !
Define your name and year:
The methodology.
What is the dataset ?
From any given movie by a user, recommendations are made from within the list of Sight and Sound top 1000 movies.
How was it done ?
The goal was to use the language model DistillBERT to perform movie recommendations. The title ans summuary of each movie is embedded using DistillBERT. Now, when we a query movie is given, we also embed it using both title and summary and request for its Nearest Neighbors within the dataset.
How to use it ?
The code has been pushed on github. Choose a movie name along with its movie year (check the correct name on wikipedia).
- As a Python package: pip install recommendation1000Movies.
- Use the function set up on Cloud function by simply running:
gcloud functions call recommending_movies --data "{\"movie_name\":\"The Big Lebowski\", \"movie_year\":\"1998\"}" --region "europe-west1"
or
curl -X POST "https://europe-west1-website-cine.cloudfunctions.net/recommending_movies" -H "Content-Type:application/json" --data "{\"movie_name\":\"The Big Lebowski\", \"movie_year\":\"1998\"}"