Offline speech recognition API for Android, iOS, Raspberry Pi and servers with Python, Java, C# and Node
Small updates to model loading
Fixes in lattice construction Transcriber with a server
Transcriber tool and model autodownload API
Add support for incremental decoding and partial lattices
Wrap exceptions on C++ size and raise them properly Fbank support Improved rescoring accuracy
Fixed RNNLM rescoring strategy, good with en-us-0.21 model.
Make word times optional. API change unfortunately.
RNNLM rescoring for big models
Different fixes for DLL loading
Move to CFFFI