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 2021/01/27 00:24:00 UTC

[jira] [Created] (MADLIB-1464) Fix fit_multiple when output_table or mst_table is passed as NULL

Nikhil Kak created MADLIB-1464:
----------------------------------

             Summary: Fix fit_multiple when output_table or mst_table is passed as NULL
                 Key: MADLIB-1464
                 URL: https://issues.apache.org/jira/browse/MADLIB-1464
             Project: Apache MADlib
          Issue Type: Bug
          Components: Deep Learning
            Reporter: Nikhil Kak
             Fix For: v1.18.0


Currently while calling fit_multiple, if we pass NULL to either the output_table argument or the mst_table argument, it fails with the wrong error message.

NULL output_table
```
select madlib.madlib_keras_fit_multiple_model(
'iris_data_packed',
NULL, 
'mst_table',
2, FALSE, NULL, 1);
ERROR:  AttributeError: FitMultipleModel instance has no attribute 'model_info_table'
```

NULL mst_table
```
select madlib.madlib_keras_fit_multiple_model(
'iris_data_packed',
'iris_multiple_model', 
NULL,
2, FALSE, NULL, 1);
ERROR:  AttributeError: FitMultipleModel instance has no attribute 'model_selection_summary_table' (plpy_elog.c:121)
```

The model hopper refactor PR https://github.com/apache/madlib/pull/525 has a fix for the NULL output_table (but the error message isn't consistent with other modules) but the null mst_table failure still exists.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)