BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks
Surat Teerapittayanon, Bradley McDanel, H. T. Kung · ICPR 2016 · 2016
- Status
- To read
- Topics
- budgeted-inference
- Feeds
- P2 · NeurIPS 2027
What they did
Adds side-branch classifiers to a deep network so that confidently classified inputs exit early, reducing inference latency and energy.
My note
No note yet.
Code
- Repository
- kunglab/branchynet
- Framework
- Chainer
- Licence
- See the repository
- My status
- Not tried
BibTeX
@article{teerapittayanon2016-branchynet,
title = {{BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks}},
author = {Surat Teerapittayanon and Bradley McDanel and H. T. Kung},
journal = {ICPR 2016},
year = {2016},
doi = {10.1109/ICPR.2016.7900006},
eprint = {1709.01686},
archivePrefix = {arXiv}
}