Frank McQuillan created MADLIB-1485:
---------------------------------------
Summary: DL: Unhelpful error message when loss/accuracy does not
exist
Key: MADLIB-1485
URL: https://issues.apache.org/jira/browse/MADLIB-1485
Project: Apache MADlib
Issue Type: Bug
Components: Deep Learning
Reporter: Frank McQuillan
Fix For: v1.19.0
We throw an unhelpful error message when either the loss or the metrics
argument is invalid. We should check if this happens for any other argument
{code}
SELECT madlib.madlib_keras_fit('mnist_train_batched', 'mnist_lstm_model',
'model_arch_mnist_lstm', 1,
$$ optimizer=SGD(lr=0.01, decay=1e-6, nesterov=True), loss='does_not_exist',
metrics=['accuracy']$$::text,
$$ batch_size=25, epochs=1, verbose=0 $$::text, 5, TRUE, NULL, NULL);
ERROR: plpy.Error: Object table not specified for function ['does_not_exist']
in compile_params (plpy_elog.c:121)
{code}
{code}
SELECT madlib.madlib_keras_fit('mnist_train_batched', 'mnist_lstm_model',
'model_arch_mnist_lstm', 1,
$$ optimizer=SGD(lr=0.01, decay=1e-6, nesterov=True),
loss='categorical_crossentropy', metrics=['foobar']$$::text,
$$ batch_size=25, epochs=1, verbose=0 $$::text, 5, TRUE, NULL, NULL);
ERROR: plpy.Error: Object table not specified for function ['foobar'] in
compile_params (plpy_elog.c:121)
{code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
