SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2211.12933 · 2022

Join the High Accuracy Club on ImageNet with A Binary Neural Network Ticket

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 8 functions out of this paper's own repositories and ran 6 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
hpi-xnor/bnext canonical 6 of 8
FunctionStatusWhere it lives
accuracy Ran hpi-xnor/bnext/utils/utils.py
code served (permissive licence) · get_code("655597f0025b0542")
adjust_temperature Ran hpi-xnor/bnext/src/train_assistant_group_amp.py
code served (permissive licence) · get_code("348e7ae7da0c6d25")
build_transform Ran hpi-xnor/bnext/utils/utils.py
code served (permissive licence) · get_code("fdd04ae393554ffa")
conv1x1 Ran hpi-xnor/bnext/src/bnext.py
code served (permissive licence) · get_code("d9def42110729a85")
conv1x1 Ran hpi-xnor/bnext/src/birealnet.py
code served (permissive licence) · get_code("f0b122e067f4f29b")
conv3x3 Ran hpi-xnor/bnext/src/bnext.py
code served (permissive licence) · get_code("71541b9507f823af")
adjust_sparse_rate Not yet run hpi-xnor/bnext/src/train_assistant_group_amp.py
code served (permissive licence) · get_code("23a8d2d1465d8757")
otsu_loss Not yet run hpi-xnor/bnext/src/train_assistant_group_amp.py
code served (permissive licence) · get_code("b1748d851bc3ef77")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Binary neural networks are the extreme case of network quantization, which has long been thought of as a potential edge machine learning solution. However, the significant accuracy gap to the full-precision counterparts restricts their creative potential for mobile applications. In this work, we revisit the potential of binary neural networks and focus on a compelling but unanswered problem: how can a binary neural network achieve the crucial accuracy level (e.g., 80%) on ILSVRC-2012 ImageNet? We achieve this goal by enhancing the optimization process from three complementary perspectives: (1) We design a novel binary architecture BNext based on a comprehensive study of binary architectures and their optimization process. (2) We propose a novel knowledge-distillation technique to alleviate the counter-intuitive overfitting problem observed when attempting to train extremely accurate binary models. (3) We analyze the data augmentation pipeline for binary networks and modernize it with up-to-date techniques from full-precision models. The evaluation results on ImageNet show that BNext, for the first time, pushes the binary model accuracy boundary to 80.57% and significantly outperforms all the existing binary networks. Code and trained models are available at: https://github.com/hpi-xnor/BNext.git.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2211.12933")
get_code_for_paper("2211.12933")
have("2211.12933")

Connect an agent — have() is free.