logo Training

TinyML and Efficient Deep Learning Computing | MIT 6.S965 Fall 2022

Total of 41 lessons ,1 days and 19 hours and 37 minutes and 26 seconds
Profile

Have you found it difficult to deploy neural networks on mobile devices and IoT devices? Have you ever found it too slow to train neural networks? This course is a deep dive into efficient machine learning techniques that enable powerful deep learning applications on resource-constrained devices. Topics cover efficient inference techniques, including model compression, pruning, quantization, neural architecture search, and distillation; and efficient training techniques, including gradient compression and on-device transfer learning; followed by application-specific model optimization techniques for videos, point cloud, and NLP; and efficient quantum machine learning. Students will get hands-on experience implementing deep learning applications on microcontrollers, mobile phones, and quantum machines with an open-ended design project related to mobile AI.

Chapters
Unfold

You Might Like

Recommended Posts

Freescale's latest in-vehicle infotainment system demonstration--EEWORLD University Hall
Freescale's latest in-vehicle infotainment system demonstration : https://training.eeworld.com.cn/course/2218
chenyy Automotive Electronics
Regarding config.bib memory configuration issue?
In the config.bib file: NK 80200000 00C00000 RAMIMAGE ;for 48M appRAM RAM 80E00000 03000000 RAM [color=#FF0000]FLASH 92000000 00100000 RESERVED[/color] What is the reserved FLASH memory segment used f
mysunshine Embedded System
PCB production via cover oil and via window
[align=left]Regarding the point of "via cover oil" and "via window", many customers often ask what this means when placing an order. Now, Jieduobang explains this issue as follows: [/align][align=left
jdbzhl PCB Design
Simulating and emulating a ZigBee wireless transmitter using ADMS RF
Today's high-tech market requires more capacity and more reliable performance of wireless technology, especially low-cost and low-power wireless communication with relatively low data rates. These typ
lorant RF/Wirelessly
Sound source localization code
[i=s]This post was last edited by littleshrimp on 2017-10-19 12:28[/i] [font=微软雅黑][size=5]Complete the SensorTile sound source localization code. All calculation codes are available. STM32L476 calcula
littleshrimp MEMS sensors
A strange problem of TMS320F2812 caused by an eCAN bug
Recently, when I was adjusting F2812 to implement the CANOpen protocol, I found a strange problem. When reading the CAN message data registers MDL and MDH of F2812, the reading was 0, as shown below.
Aguilera DSP and ARM Processors
Recommended Content
Web users are watching Change

EEWorld
subscription
account

EEWorld
service
account

Automotive
development
circle

About Us Customer Service Contact Information Datasheet Sitemap LatestNews


Room 1530, 15th Floor, Building B, No.18 Zhongguancun Street, Haidian District, Beijing, Postal Code: 100190 China Telephone: 008610 8235 0740

Copyright © 2005-2024 EEWORLD.com.cn, Inc. All rights reserved 京B2-20211791 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号