kaknikhil commented on a change in pull request #432: MADLIB-1351 : Added
stopping criteria on perplexity to LDA
URL: https://github.com/apache/madlib/pull/432#discussion_r320479364
##########
File path: src/ports/postgres/modules/lda/lda.py_in
##########
@@ -445,6 +511,12 @@ def lda_train(schema_madlib, train_table, model_table,
output_data_table, voc_si
'invalid argument: positive real expected for alpha')
_assert(beta is not None and beta > 0,
'invalid argument: positive real expected for beta')
+ _assert(evaluate_every is not None and evaluate_every >= 0,
Review comment:
The user docs for evaluate_every mention `Set it to 0 or negative number to
not evaluate perplexity in training at all` but this check will throw an
exception for evaluate_every < 0
----------------------------------------------------------------
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
