Advancing Automatic Speech Recognition with Edge Computing

As the world becomes increasingly digital, conversational AI has become a common way to achieve interaction between humans and computers. Nemo is designed for developers curious about “conversational AI”; it is an open-source toolkit based on PyTorch that allows developers to quickly build models for real-time Automatic Speech Recognition (ASR), Natural Language Processing (NLP), and Text-to-Speech (TTS) applications. Conversational AI shapes the path of human-computer interaction, making it more accessible and helping to bridge the gap between machines and humans.

Previously, the majority of AI relied on cloud computing, as the cloud offers richer computational power, GPU resources, and machine learning platforms. With the continuous development of AI chips and edge computing power, Edge Intelligence (EI) has become a trend for the future, and more people are choosing to deploy AI on “edge computing devices”.

Since its launch in 2019, the NVIDIA Jetson Nano has sparked a wave of interest in the global AIOT edge computing application field, even winning the “Best AI Processor” award in the 2020 Best Vision Products list published by the Edge AI and Vision Alliance.

So, how can you deploy Nemo-trained automatic speech models on the Jetson Nano? How can you utilize Nemo on the Jetson Nano?

NVIDIA, in collaboration with InfoQ, provides a series of online training courses related to AI development for developers with high-performance computing and artificial intelligence development needs, breaking down barriers and helping you get started quickly.

In the previous session, NVIDIA Developer Community Manager Li Yipeng introduced the workflow and system architecture of ASR, detailing the pre-trained ASR model Quartznet, guiding viewers into the basics of using Nemo to quickly complete transfer learning tasks in automatic speech recognition.

On April 28, 2021, from 20:00 to 21:30, Li Yipeng will present the fifth session of the public course—quickly implementing ASR applications on edge computing devices.

This online seminar is primarily aimed at developers with needs in speech semantics and artificial intelligence development. Through this online seminar, you can gain the following content:

  • Introduction to Jetson Nano and the conversational AI toolkit NeMo

  • Learn to set up the prerequisites for NeMo installation

  • Installation guide for NeMo on Jetson Nano

  • Completing Chinese speech recognition tasks with NeMo on Jetson Nano

  • Deploying the trained model on Jetson Nano for inference

……

Scan the QR code below or click 【Read Original】 to register for free.

Advancing Automatic Speech Recognition with Edge Computing

Tap to view fewer bugs 👇

Leave a Comment