</> Code with Papers by Syntology

Papers › SCARLET-NAS: Bridging the Gap between Stability and Scalability in Weight-sharing...

SCARLET-NAS: Bridging the Gap between Stability and Scalability in Weight-sharing Neural Architecture Search

16 Aug 2019arXiv:1908.06022archive 2025-07-28

Xiangxiang Chu, Bo Zhang, Qingyuan Li, Ruijun Xu, Xudong Li

To discover powerful yet compact models is an important goal of neural architecture search. Previous two-stage one-shot approaches are limited by search space with a fixed depth. It seems handy to include an additional skip connection in the search space to make depths variable. However, it creates a large range of perturbation during supernet training and it has difficulty giving a confident ranking for subnetworks. In this paper, we discover that skip connections bring about significant feature inconsistency compared with other operations, which potentially degrades the supernet performance. Based on this observation, we tackle the problem by imposing an equivariant learnable stabilizer to homogenize such disparities. Experiments show that our proposed stabilizer helps to improve the supernet's convergence as well as ranking performance. With an evolutionary search backend that incorporates the stabilized supernet as an evaluator, we derive a family of state-of-the-art architectures, the SCARLET series of several depths, especially SCARLET-A obtains 76.9% top-1 accuracy on ImageNet. Code is available at https://github.com/xiaomi-automl/ScarletNAS.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

xiaomi-automl/SCARLET-NAS officialmentioned in papermentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

AutoMLImage ClassificationNeural Architecture Search

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification ImageNet SCARLET-A4 GFLOPs 8.4 #551 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-A4 Hardware Burden 12G #551 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-A4 Number of params 27.8M #551 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-A4 Operations per network pass 0.42G #551 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-A4 Top 1 Accuracy 82.3% #551 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-A GFLOPs 0.730 #895 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-A Number of params 6.7M #895 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-A Top 1 Accuracy 76.9% #895 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-B GFLOPs 0.658 #919 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-B Number of params 6.5M #919 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-B Top 1 Accuracy 76.3% #919 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-C GFLOPs 0.560 #945 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-C Number of params 6M #945 of 1060 Archive leaderboard report
Image Classification ImageNet SCARLET-C Top 1 Accuracy 75.6% #945 of 1060 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-A Accuracy 76.9 #76 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-A MACs 365M #76 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-A Params 6.7M #76 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-A Top-1 Error Rate 23.1 #76 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-B Accuracy 76.3 #86 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-B MACs 329M #86 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-B Params 6.5M #86 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-B Top-1 Error Rate 23.7 #86 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-C Accuracy 75.6 #103 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-C MACs 280M #103 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-C Params 6.0M #103 of 135 Archive leaderboard report
Neural Architecture Search ImageNet SCARLET-C Top-1 Error Rate 24.4 #103 of 135 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Introduced by this paper: SCARLET-NAS

1x1 ConvolutionAutoAugmentBatch NormalizationColorJitterCosine AnnealingCutoutDense ConnectionsDepthwise ConvolutionDepthwise Separable ConvolutionDropoutInverted Residual BlockLSTMPointwise ConvolutionRMSPropResidual ConnectionSCARLETSCARLET-NASSigmoid ActivationStep DecayTanh ActivationWeight Decay

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections