--picklefile

Switch

--picklefile

Description

Switch to indicate name of pickle file to save/load a model from.

Argument and Default Value

Name of picklefile

Details

If you are running a command that creates/trains or loads/uses a model, using this will tell the infrastructure where to find/put the model. This switch requires the use of --load_model or --save_model.

Other Switches

Required Switches: -d, -g, -t, -f, --outcome_table, --outcomes --save_model or --load_model Example Commands ================

Using the packaged dla_tutorial dataset (see Building A Pickle Model and Applying A Pickle Model):

# Trains a regression model to predict age for users from 1grams
# Will save the model to a picklefile called age.pickle
dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id$16to16$0_01' \
--outcome_table blog_outcomes --outcomes age --train_regression --save_model --picklefile age.pickle

# Uses the trained regression model (age.pickle) to predict age for users from 1grams
dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id$16to16$0_01' \
--outcome_table blog_outcomes --outcomes age --predict_regression --load_model --picklefile age.pickle