-
Notifications
You must be signed in to change notification settings - Fork 750
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Mqcnn rts #668
Mqcnn rts #668
Conversation
Codecov Report
@@ Coverage Diff @@
## master #668 +/- ##
=======================================
Coverage 85.87% 85.87%
=======================================
Files 172 172
Lines 11037 11037
=======================================
Hits 9478 9478
Misses 1559 1559 |
The option to ignore dynamical-features will need some refactoring in the |
So from master there was a change that removed shuffling all-together except on a single batch basis. |
decoder_input_static, _, decoder_input_dynamic = self.enc2dec( | ||
decoder_input_static, decoder_input_dynamic = self.enc2dec( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This must have been a pretty big bug in the seq2seq
library right? If you look at the default enc2dec
the third second output is the decoder_input_dynamic, not the third, thus decoder_input_dynamic has been dropped thus far and future_dynamic_feat used in its place
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I also removed the third output of enc2dec
because it was never used and I did not see the point, but I'm open to putting that back.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@lostella do you have any preference here regarding my last comment?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good to me! What an effort! Any last-minute adjustments? After those this can go in
Issue #, if available:
Description of changes:
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.
Todo
add support for categorial features
option to ignore dynamical-features (like
ignore
in SageMaker DeepAR)fix instance splitter
add observed value indicator (see Add missing value support for ForkingSeq2Seq Estimators (MQCNN, MQRNN) #704)