Riva Speech Recognition (SR) Plugin 1.0.0 to the UniMRCP Server (UMS) has been released.
The plugin is based on the following components:
- UniMRCP Server 1.7.0
- NVIDIA Riva Speech Skills 1.9.0-beta
- gRPC 1.30.3
- Protobuf 3.12.2
The binaries are currently available for the following Linux distributions:
- Red Hat / CentOS 7 (unimrcp-riva-sr-1.7.0-1.el7.x86_64.rpm)
- Red Hat / CentOS 8 (unimrcp-riva-sr-1.7.0-1.el8.x86_64.rpm)
- Ubuntu 18.04 LTS (unimrcp-riva-sr_1.7.0-bionic_amd64.deb)
- Ubuntu 20.04 LTS (unimrcp-riva-sr_1.7.0-focal_amd64.deb)
IVR platforms can now utilize NVIDIA Riva Speech-to-Text API via UniMRCP Server.
NVIDIA Riva Speech-to-Text API performs speech to text conversion powered by machine learning providing the following main features.
Streaming Speech Recognition
Supports efficient streaming speech transcription.
Low Latency
Intermediate transcripts are returned with low latency.
Efficient Feature Extraction
GPU-accelerated feature extraction.
Multiple Acoustic Models
Multiple (and growing) acoustic model architecture options accelerated by NVIDIA TensorRT
Beam Search Decoder
Beam search decoder based on n-gram language models
Voice Activity Detection
CTC-based voice activity detection algorithms.
Automatic Punctuation
Automatic punctuation can optionally be enabled.
Alternate Transcripts
Ability to return top-N transcripts from beam decoder
Word-level Timestamps
Word-level timestamps can optionally be returned.
Inverse Text Normalization
Inverse text normalization (ITN) is supported.
Thank you for using UniMRCP.
--
Arsen Chaloyan
Author of UniMRCP
http://www.unimrcp.org
Downloads
Repository
Documentation
Issues
Project Media
Google Cloud
Microsoft Azure
Amazon
Nuance Mix
IBM Watson
Yandex Cloud
NVIDIA
SoundHound
Verbio
Deepgram
ASAPP
Speechmatics
GoVivace
Voci
Kaldi
Platforms
Announcement Group