The speech confabulation model is built based on cogent confabulation to process streaming input data to recover a sentence from speech wav. The speech wav is handled using GMM model to give likelihood of each phone as input of confabulation. To improve the performance, sentence level confabulation is introduced.
-
Notifications
You must be signed in to change notification settings - Fork 0
License
lz1313/speech
About
No description, website, or topics provided.
Resources
License
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published