Главная
Study mode:
on
1
Introduction
2
Introducing Seafood
3
Welcome
4
Why do we care
5
Latency
6
Push notifications
7
Training a model
8
Finetuning
9
Tensorflow
10
Hardware
11
Energy Considerations
12
Model Pruning
13
Quantization
14
Pocket Flow
15
Neural Architecture Search
16
Training Time
17
Comparison
18
Ondevice training
19
federated learning
20
recap
21
book
22
Questions
Description:
Explore the challenges and solutions for implementing deep learning on mobile devices in this 44-minute conference talk. Learn about convolutional neural networks (CNNs) and their potential applications in smartphone and wearable device applications. Discover strategies to overcome memory and power constraints, including building mobile-friendly shallow CNN architectures. Gain insights from real-time demos and case studies from major tech companies like Google, Microsoft, and Facebook. Delve into topics such as model pruning, quantization, neural architecture search, on-device training, and federated learning. Understand the importance of latency, energy considerations, and hardware limitations in mobile deep learning implementations. Walk away with practical knowledge to leverage deep learning techniques in resource-constrained mobile environments.

Deep Learning for Mobile Devices

WeAreDevelopers
Add to list
0:00 / 0:00