Sharpness-Aware Minimization for Efficiently Improving Generalization

Code

• github.com/google-research/sam
  OfficialIn paperjax★ 624
• github.com/davda54/sam
  pytorch★ 1,966
• github.com/moskomule/sam.pytorch
  pytorch★ 135
• github.com/simon20010923/DDAMFN
  pytorch★ 116
• github.com/ys-zong/medfair
  pytorch★ 73
• github.com/sayakpaul/Sharpness-Aware-Minimization-TensorFlow
  tf★ 61
• github.com/wangermeng2021/Scaled-YOLOv4-tensorflow2
  tf★ 47
• github.com/Jannoshh/simple-sam
  tf★ 41
• github.com/rollovd/LookSAM
  pytorch★ 37
• github.com/wangermeng2021/FastClassification
  tf★ 15

Abstract

In today's heavily overparameterized models, the value of the training loss provides few guarantees on model generalization ability. Indeed, optimizing only the training loss value, as is commonly done, can easily lead to suboptimal model quality. Motivated by prior work connecting the geometry of the loss landscape and generalization, we introduce a novel, effective procedure for instead simultaneously minimizing loss value and loss sharpness. In particular, our procedure, Sharpness-Aware Minimization (SAM), seeks parameters that lie in neighborhoods having uniformly low loss; this formulation results in a min-max optimization problem on which gradient descent can be performed efficiently. We present empirical results showing that SAM improves model generalization across a variety of benchmark datasets (e.g., CIFAR-10, CIFAR-100, ImageNet, finetuning tasks) and models, yielding novel state-of-the-art performance for several. Additionally, we find that SAM natively provides robustness to label noise on par with that provided by state-of-the-art procedures that specifically target learning with noisy labels. We open source our code at https://github.com/google-research/sam.

Tasks

Benchmark Results

Reproductions