Turicreate: Departure from Graphlab#

Turi Create simplifies the development of custom machine learning models.

apple/turicreate

You don’t have to be a machine learning expert to add

  • recommendations,

  • object detection,

  • image classification,

  • image similarity

  • activity classification

to your app.

pip install -U turicreate#

import turicreate as tc
actions = tc.SFrame.read_csv('/Users/datalab/bigdata/cjc/ml-1m/ratings.dat', delimiter='\n', 
                                header=False)['X1'].apply(lambda x: x.split('::')).unpack()
for col in actions.column_names():
    actions[col] = actions[col].astype(int)
actions = actions.rename({'X.0': 'user_id', 'X.1': 'movie_id', 'X.2': 'rating', 'X.3': 'timestamp'})
#actions.save('ratings')

Finished parsing file /Users/datalab/bigdata/cjc/ml-1m/ratings.dat
Parsing completed. Parsed 100 lines in 0.348264 secs.
------------------------------------------------------
Inferred types from first 100 line(s) of file as 
column_type_hints=[str]
If parsing fails due to incorrect types, you can correct
the inferred type list above and pass it to read_csv in
the column_type_hints argument
------------------------------------------------------

Finished parsing file /Users/datalab/bigdata/cjc/ml-1m/ratings.dat
Parsing completed. Parsed 1000209 lines in 0.376001 secs.