[
https://issues.apache.org/jira/browse/MADLIB-1349?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16851390#comment-16851390
]
Frank McQuillan commented on MADLIB-1349:
-----------------------------------------
LGTM
See the tests at the bottom of
https://issues.apache.org/jira/browse/MADLIB-1338 showing selection of
different metrics. Regarding change of model size to KB, this seems to be
fixed too:
{code}
madlib=# select * from iris_model_summary;
-[ RECORD 1
]-------------+--------------------------------------------------------------------------
source_table | iris_train_packed
model | iris_model
dependent_varname | class_text
independent_varname | attributes
model_arch_table | model_arch_library
model_arch_id | 1
compile_params | loss='categorical_crossentropy', optimizer='adam',
metrics=['accuracy']
fit_params | batch_size=16, epochs=1
num_iterations | 3
validation_table | iris_test_packed
metrics_compute_frequency | 2
name | Sophia L.
description | Simple MLP model on iris dataset
model_type | madlib_keras
model_size | 0.7939453125
start_training_time | 2019-05-29 20:01:42.08684
end_training_time | 2019-05-29 20:01:48.658938
metrics_elapsed_time | {5.42638921737671,6.57209706306458}
madlib_version | 1.16-dev
num_classes | 3
class_values | {Iris-setosa,Iris-versicolor,Iris-virginica}
dependent_vartype | character varying
normalizing_const | 1
metrics_type | {accuracy}
training_metrics_final | 0.341666668653
training_loss_final | 1.35804212093
training_metrics | {0.349999994039536,0.341666668653488}
training_loss | {1.42269325256348,1.35804212093353}
validation_metrics_final | 0.266666680574
validation_loss_final | 1.50677502155
validation_metrics | {0.266666680574417,0.266666680574417}
validation_loss | {1.58999609947205,1.50677502155304}
metrics_iters | {2,3}
{code}
> DL: Add metric_type and fix model_size in model summary table
> -------------------------------------------------------------
>
> Key: MADLIB-1349
> URL: https://issues.apache.org/jira/browse/MADLIB-1349
> Project: Apache MADlib
> Issue Type: New Feature
> Components: Deep Learning
> Reporter: Nandish Jayaram
> Priority: Major
> Fix For: v1.16
>
>
> We plan to support only one metric in 1.16 (
> https://github.com/apache/madlib/pull/393 ) and that can be one of the
> supported metrics and not just accuracy. We should store the name of the
> metric that was computed (if computed) in the summary table.
> There also seems to be a bug in reporting the {{model_size}} in the summary
> table, as it reports a drastically small number compared to the actual model
> learnt. Since that is also an update to summary table, it may be a good idea
> to fix it as part of this JIRA.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
