Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Benchmark: Layer-wise training of VGG-like network: 95.86% test accuracy #138

Open
anokland opened this issue Jan 21, 2019 · 0 comments
Open

Benchmark: Layer-wise training of VGG-like network: 95.86% test accuracy #138

anokland opened this issue Jan 21, 2019 · 0 comments

Comments

@anokland
Copy link

anokland commented Jan 21, 2019 •
edited

A VGG-like network with 6 convolutional layers and 1 fully connected layer.

conv256-conv512-maxpool-conv512-conv1024-maxpool-conv1024-maxpool-conv1024-maxpool-fc1024-fc10

  • Standard preprocessing (mean/std subtraction/division)
  • Data augmentation (random crops/horizontal flips)
  • Cutout
  • 28M parameters
  • Layer-wise training, no global back-propagation
  • Code and more results: https://github.com/anokland/local-loss
@anokland anokland changed the title Benchmark: Layer-wise training of VGG-like network: 95.86% test error Benchmark: Layer-wise training of VGG-like network: 95.86% test accuracy Jan 21, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@anokland