This repository has been archived by the owner on Oct 13, 2021. It is now read-only.
Support initial states for Bidirectional RNN #417
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…cover RNN initial state passing
|
This pull request fixes 2 alerts when merging ffb7041 into aad8ed8 - view on LGTM.com fixed alerts:
|
|
Looks like there are issues in Python 3.7 that need to be addressed. |
|
This pull request fixes 2 alerts when merging 9cc6446 into aad8ed8 - view on LGTM.com fixed alerts:
|
|
@wenbingl, can you review this PR? |
wenbingl
approved these changes
Mar 30, 2020
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR provides support for initial states in Bidirectional SimpleRNN, GRU, and LSTM layers. It also fixes a bug where leveraging two outputs of a Keras node can cause parsing failures (see the changes in
parser.py). Support was extended for LSTM and Bidirectional layers that have dynamic sequence lengths (seq_length = None), so that it now supports down to opset 5. This was previously possible in terms of the ONNX spec, but would raise an error.