You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@madlib.apache.org by "Nikhil Kak (Jira)" <ji...@apache.org> on 2021/03/19 20:22:00 UTC
[jira] [Commented] (MADLIB-1470) mlp_classification and weights
[ https://issues.apache.org/jira/browse/MADLIB-1470?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17305175#comment-17305175 ]
Nikhil Kak commented on MADLIB-1470:
------------------------------------
[~kevins1101A]
Thanks for reporting the issue. This should now be fixed as part of https://github.com/apache/madlib/pull/557. The fix will be released as part of the next madlib release 1.18.0. If you want to try it before the release, you can compile madlib from source.
> mlp_classification and weights
> ------------------------------
>
> Key: MADLIB-1470
> URL: https://issues.apache.org/jira/browse/MADLIB-1470
> Project: Apache MADlib
> Issue Type: Question
> Components: Module: Neural Networks
> Reporter: Kevin
> Priority: Major
>
> Trying to use mlp_classification with Greenplum
> Doc states weights should be numeric column
> Have tried following specifications (column name specified: 'target_weight')
> target_weight numeric(5,3)
> target_weight numeric(4,3)
>
> but I get following error when attempting to use function mlp_classification with above specification:
> ERROR: plpy.Error: MLP error: Weights should be a numeric type (plpy_elog.c:121)
> CONTEXT: Traceback (most recent call last):
> PL/Python function "mlp_classification", line 33, in <module>
> grouping_col)
> PL/Python function "mlp_classification", line 42, in wrapper
> PL/Python function "mlp_classification", line 116, in mlp
> PL/Python function "mlp_classification", line 740, in _validate_params_based_on_minibatch
> PL/Python function "mlp_classification", line 123, in _assert
> PL/Python function "mlp_classification"
--
This message was sent by Atlassian Jira
(v8.3.4#803005)