How to predict for new users?

[levalencia]

Great package, I was able to run it and do a prediction for an existing user:

scores = model.predict(user_ids=6, item_ids=[1,2,3,5,6])
print(scores)

However, I would like to know how to make predictions for new users (cold start), I am not able to find documentation here or on light fm about how do it.

I tried this

new_user_feature = [8,{'name:John', 'Age:33', 'los:IFS','ou:development', 'skills:sql'} ]    
new_user_feature = [8,new_user_feature]

#predict new users User-Id	name	age	los	ou	gender	skills
model.predict(0, item_ids=[1,2,3,5,6], user_features=new_user_feature)

But I get this error:

<1 sec
AttributeError: 'list' object has no attribute 'tocsr'

any idea?