-
Notifications
You must be signed in to change notification settings - Fork 16
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
Generic Seq2Seq Decoder preparation #49
Conversation
@@ -37,10 +39,11 @@ namespace Speech { | |||
class ModelCombination : public Mc::Component, public Core::ReferenceCounted { | |||
public: | |||
typedef u32 Mode; | |||
static const Mode complete; | |||
static const Mode complete; // Includes lexicon, AM and LM but NOT label scorer; named 'complete' for legacy reasons. |
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.
While the name complete
is a bit misleading now because it doesn't include the label scorer, changing it would mean a config-level change since the name 'complete' is also used in src/Speech/LatticeNodes.cc
for a Core::Choice ModelCombinationNode::choiceMode
. So a comment was added here and the name is left as-is for now.
This PR contains a subset of changes from #45. These are all the smaller file changes that prepare the rest of the code base for the new module but omitting actual implementations of the new label scorers and search/alignment algorithms.
The code in this PR can be compiled successfully but the new classes are mostly empty and non-functional at this stage.