njayaram2 opened a new pull request #360: Deep Learning: Add support for one-hot encoded dep var URL: https://github.com/apache/madlib/pull/360 The current madlib_keras_fit function for DL assumes dependent variable is not one-hot encoded. But fit uses data obtained after running minibatch_preprocessor_dl that returns a 1-hot encoded array for each dependent var (https://issues.apache.org/jira/browse/MADLIB-1303). This commit updates madlib_keras_fit to expect a 1-hot encoded dependent var to train the model. We also create a new column named class_values in the model summary table that is passed along from the input summary table (which is mini-batched). Predict then uses class_values to find the exact class label for test data. class_values can be NULL in case where user had already 1-hot encoded the data before calling minibatch_preprocessor_dl. In that case, we return the index of the expected class label.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
