njayaram2 commented on a change in pull request #397: DL: Update content of
the model summary table
URL: https://github.com/apache/madlib/pull/397#discussion_r288342302
##########
File path: src/ports/postgres/modules/deep_learning/madlib_keras.py_in
##########
@@ -85,14 +66,17 @@ def fit(schema_madlib, source_table,
model,model_arch_table,
independent_varname = MINIBATCH_OUTPUT_INDEPENDENT_COLNAME_DL
model_arch_table = quote_ident(model_arch_table)
fit_params = "" if not fit_params else fit_params
+ compile_params = "" if not compile_params else compile_params
Review comment:
This does not, but compile_params is quoted using `$madlib$` before passing
it along to the transition function. We cannot do that to `None`, so this only
changes None to an empty string. The actual validation happens in
`parse_and_validate_compile_params()` in `madlib_keras_wrapper.py_in`.
In the current state of the code, compile params are validated only in
`fit_transition`. That's not ideal, but I think there is some other JIRA to
refactor that code and ensure validation happens in `fit`.
----------------------------------------------------------------
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
