NVIDIA Jarvis Speech Skills
v1.0.0-b.2
Version select:

About

  • Quick Start Guide
    • Prerequisites
    • Models Available for Deployment
    • Local Deployment using Quick Start Scripts
    • Running the Jarvis Client and Transcribing Audio Files
    • Running the Jarvis Client and Converting Text to Audio Files
  • Release Notes
    • Jarvis Speech Skills 1.0.0-b.2 Beta
      • Key Features and Enhancements
      • Bugfixes
      • Compatibility
      • Limitations
    • Jarvis Speech Skills 1.0.0-b.1 Beta
      • Key Features and Enhancements
      • Compatibility
      • Known Issues
      • Limitations
  • Support Matrix
    • Jarvis Speech Skills 1.0.0 Beta
      • Hardware
      • Software
  • Best Practices
    • Using Your Own Data
    • Autoscaling Configurations
    • gRPC Streams with SSL/TLS
    • Monitoring GPU and CPU Telemetry
    • Load Balancing Types

Services and Models

  • Overview
    • Model Development with NeMo
    • Model Development with TLT
      • TLT Export for NeMo/TLT
    • Model Development with NeMo
    • Jarvis Build
    • Jarvis Deploy
    • NeMo to Jarvis
  • Speech Recognition
    • Model Architectures
      • Jasper
      • QuartzNet
    • Services
      • Offline Recognition
      • Streaming Recognition
    • Pipeline Configuration
      • Streaming/Offline Configuration
      • Language Models
      • GPU-accelerated Decoder
      • Beginning/End of Utterance Detection
      • Non-English Languages
      • Selecting Custom Model at Runtime
    • Pretrained Models
  • Natural Language Processing
    • Model Architectures
      • Bidirectional Encoder Representations from Transformers (BERT)
      • Megatron
    • Services
    • Pipeline Configuration
      • Text Classification
      • Token Classification (Named Entity Recognition)
      • Joint Intent and Slots
      • Question Answering (Extractive)
      • Punctuation and Capitalization
    • Pretrained Models
  • Speech Synthesis
    • Model Architectures
    • Services
    • Pipeline Configuration
    • Pretrained Models
  • Deploying Your Custom Model into Jarvis
    • Build process
    • Deploy process
      • Option 1: Using Quick Start Scripts to Deploy Your Models (Recommended path)
      • Option 2: Using jarvis-deploy and the Jarvis Speech Container (Advanced)

Client Integration

  • gRPC API
    • Generating Bindings
    • Tutorial
  • Python API
    • Installation
    • Python Example
  • Command-line Clients
    • Speech Recognition
      • C++ Streaming Example
      • C++ Offline/Batch (non-streaming) Example
      • Python Streaming Example
    • Natural Language Processing
      • C++ NER Example
      • C++ QA Example
      • Python QA Example
      • C++ Punctuation Example
    • Speech Synthesis
      • C++ TTS Client Example
      • C++ TTS Performance Client Example
      • Python Client Examples

Server Deployment

  • Local (Docker)
    • Configuring
    • Downloading Required Models and Containers from NGC
    • Launching the Servers and Client Container
    • Stopping
    • Clean-up
  • Kubernetes
    • Required Software
    • Kubernetes Secrets
    • Jarvis Settings
    • Ingress Controller
    • Load Balancer
  • NVIDIA Fleet Command
    • Application Setup
    • Deployment Setup
  • Performance
    • ASR
      • NVIDIA A100 GPU
        • 100ms chunk
        • 800ms chunk
        • 3200ms chunk
      • NVIDIA V100 GPU
        • 100ms chunk
        • 800ms chunk
        • 3200ms chunk
      • NVIDIA T4
        • 100ms chunk
        • 800ms chunk
        • 3200ms chunk
    • NLP
      • NVIDIA A100 GPU
      • NVIDIA V100 GPU
      • NVIDIA T4
    • TTS
      • NVIDIA A100 GPU
      • NVIDIA V100 GPU
      • NVIDIA T4

Reference

  • gRPC & Protocol Buffers
    • src/jarvis_proto/jarvis_asr.proto
    • src/jarvis_proto/jarvis_nlp.proto
    • src/jarvis_proto/jarvis_nlp_core.proto
    • src/jarvis_proto/jarvis_tts.proto
    • src/jarvis_proto/audio.proto
    • src/jarvis_proto/health.proto
  • NGC Artifacts
    • Notebooks
    • Models
    • NGC Collections
  • Troubleshooting
    • NGC
    • Model Export and ServiceMaker
    • Services
    • Client Integration
    • Jarvis Helm Deployment
    • SpeechSquad
  • Acknowledgements
    • Google APIs
    • GoogleTest
    • gflags
    • Google Logging Library (glog)
    • speexdsp
    • libFLAC
    • gRPC
    • Triton Inference Server
    • NVlabs cub
    • KenLM
    • Kaldi
    • grpc_health_probe
    • gRPCurl
    • OpenFST
    • Yamale
    • PyTorch
    • requests
    • PyCUDA
    • RapidJSON
    • protobuf
    • onnx
    • librosa
    • omegaconf

Sample Applications

  • SpeechSquad
    • Local Deployment
      • Installation
      • Running the Test Locally
    • Cloud Deployment
      • Installation
      • Containers
      • Dataset
      • Running the Test
  • Virtual Assistant
    • Video Demo
    • Running the Demo
    • Sample Use Cases
    • Limitations
    • License
  • Virtual Assistant (with Rasa)
    • Demo Video
    • Implementation
      • Architecture
      • Code Structure
      • Network Configuration
    • Running the Demo
    • Sample Conversations
    • Limitations
    • License
  • Contact Center Video Conference
    • Requirements
    • Installation
    • Running the Service
    • Using the Service
  • Deploying Jarvis ASR Service on AWS EKS
    • Downloading and Modifying the Jarvis API Helm Chart
    • Downloading and Modifying the Traefik Helm Chart
    • Defining the Client Service and Ingress Route
    • Defining and Launching the EKS Cluster
    • Running the Benchmarks
    • Scaling and Deleting the Cluster

Sample Notebooks

  • Python API Examples
    • Overview
    • Introduction the Jarvis Speech and Natural Languages services
    • Learning objectives
    • Requirements and setup
      • Create Jarvis clients and connect to Jarvis Speech API server
    • Content
    • 1. Offline ASR Example
    • 2. Core NLP Service Examples
    • 3. TTS Service Example
    • 4. Jarvis NLP Service Examples
    • 5. Go deeper into Jarvis capabilities
      • 1. Sample apps:
      • 3. Finetune your own domain specific Speech or NLP model and deploy into Jarvis.
      • 3. Further resources:
  • Wikipedia Question Answering
    • Wikipedia Summary
    • Query Jarvis Server
NVIDIA Jarvis Speech Skills
  • Docs »
  • NGC Artifacts

NGC Artifacts¶

Notebooks¶

These notebooks provide an end-to-end workflow for the following services starting with training in TLT and deployment using Jarvis.

  • Speech Recognition notebook

  • Question Answering notebook

  • Text Classification notebook

  • Named Entity Recognition (NER) notebook

  • Punctuation Capitalization notebook

  • Intent Slot Classification notebook

Models¶

  • Speech Recognition English Jasper model

  • Speech Recognition English QuartzNet model

  • Question Answering SQUAD2.0 Megatron model

  • Question Answering SQUAD2.0 Bert Base model

  • Question Answering SQUAD2.0 Bert Large model

  • Domain Classification Bert model

  • Punctuation and Capitalization Bert model

  • Named Entity Recognition Bert model

  • Intent and Slot Classification Bert model

  • Speech Synthesis Tacotron 2 model

  • Speech Synthesis WaveGlow model

NGC Collections¶

  • Jarvis NGC collection

  • TLT Conversational AI collection

  • TLT/Jarvis Speech Recognition collection

  • TLT/Jarvis Question Answering collection

  • TLT/Jarvis Text Classification collection

  • TLT/Jarvis Named Entity Recognition collection

  • TLT/Jarvis Punctuation and Capitalization collection

  • TLT/Jarvis Intent Detection and Slot Tagging collection

  • Jarvis Speech Synthesis collection

Next Previous

© Copyright 2018-2021, NVIDIA Corporation

Built with Sphinx using a theme provided by Read the Docs.