On-Device Learning Under 256KB Memory - Challenges and Solutions for IoT Devices

On-Device Learning Under 256KB Memory - Challenges and Solutions for IoT Devices

tinyML via YouTube Direct link

Tiny Training Engine Workflow

10 of 13

10 of 13

Tiny Training Engine Workflow

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

On-Device Learning Under 256KB Memory - Challenges and Solutions for IoT Devices

Automatically move to the next video in the Classroom when playback concludes

  1. 1 Intro
  2. 2 Can we Learn on the Edge? Al systems need to continually adapt to new data collected from the sensors Not only inference, but also run back-propagation on edge devices
  3. 3 #Activation is the Memory Bottleneck, not #Trainable Parameters
  4. 4 Related Work: Parameter-Efficient Transfer Learning
  5. 5 Address Optimization Difficulty of Quantized Graphs
  6. 6 QAS: Quantization-Aware Scaling
  7. 7 Sparse Layer/Tensor Update
  8. 8 Find Layers to Update by Contribution Analysis
  9. 9 Tiny Training Engine (TTE)
  10. 10 Tiny Training Engine Workflow
  11. 11 Deep Gradient Compression: Reduce Bandwidth
  12. 12 Post-training testing (high accuracy)
  13. 13 Real-life testing

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.