Papers › Res2Net: A New Multi-scale Backbone Architecture

Res2Net: A New Multi-scale Backbone Architecture

2 Apr 2019arXiv:1904.01169archive 2025-07-28

Shang-Hua Gao, Ming-Ming Cheng, Kai Zhao, Xin-Yu Zhang, Ming-Hsuan Yang, Philip Torr

Representing features at multiple scales is of great importance for numerous vision tasks. Recent advances in backbone convolutional neural networks (CNNs) continually demonstrate stronger multi-scale representation ability, leading to consistent performance gains on a wide range of applications. However, most existing methods represent the multi-scale features in a layer-wise manner. In this paper, we propose a novel building block for CNNs, namely Res2Net, by constructing hierarchical residual-like connections within one single residual block. The Res2Net represents multi-scale features at a granular level and increases the range of receptive fields for each network layer. The proposed Res2Net block can be plugged into the state-of-the-art backbone CNN models, e.g., ResNet, ResNeXt, and DLA. We evaluate the Res2Net block on all these models and demonstrate consistent performance gains over baseline models on widely-used datasets, e.g., CIFAR-100 and ImageNet. Further ablation studies and experimental results on representative computer vision tasks, i.e., object detection, class activation mapping, and salient object detection, further verify the superiority of the Res2Net over the state-of-the-art baseline methods. The source code and trained models are available on https://mmcheng.net/res2net/.

PaperPDFCodeCode Syntology ran

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

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="1904.01169")

Code

Syntology Ran 3 of 9 code samples harvested from 3 repositories linked to this paper; 6 have no recorded run. Of those that ran: 2 ran · our draft was wrong; 1 ran with no contract checked.

By repository: community (archive-listed): 8 samples from 3 repositories, 2 ran; 1 identical to code first harvested elsewhere. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

34 repositories listed; official and paper-mentioned ones first.

ChrisMats/Res2Net mentioned on GitHubpytorch report
IMvision12/keras-vision-models mentioned on GitHubpytorch report
Res2Net/Res2Net-PoolNet mentioned on GitHubpytorch report
Res2Net/Res2Net-Pose-Estimation mentioned on GitHubpytorch report
Res2Net/Res2Net-PretrainedModels mentioned on GitHubpytorch report
Res2Net/Res2Net-maskrcnn mentioned on GitHubpytorchMIT report
fupiao1998/res2net-keras mentioned on GitHubtf report
gasvn/Res2Net mentioned on GitHubpytorch report
kingcong/res2net mentioned on GitHubmindsporenot reachable when probed 2026-09-18 — repositories for recent papers often appear after camera-ready report
osmr/imgclsmob mentioned on GitHubmxnetMIT report
rwightman/pytorch-image-models mentioned on GitHubpytorch report
tuanzhangCS/res2net-on-mxnet mentioned on GitHubmxnet report
yfreedomliTHU/Res2Net mentioned on GitHubpytorch report
zhuhongwei1999/bsa-net mentioned on GitHubpytorch report
PaddlePaddle/PaddleClas paddleApache-2.0 report
Res2Net/mmdetection pytorchApache-2.0 report
open-mmlab/mmdetection pytorchApache-2.0 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

9 samples harvested; 3 ran; 0 honoured the contract we drafted; 6 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

2ran · our draft was wrong
1ran
6unverified

Licence: 9 of the 9 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 3 repositories linked to this paper, official or community; each sample names its own and says which. Some samples are identical code Syntology first harvested from another repository; for those, this paper's copy is not located and its licence is not recorded. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

conv3x3 ChrisMats/Res2Net/res2net.py community (archive-listed) ran · our draft was wrong no licence file found · pointer only · e18ed73aa89f051a · report
smooth_l1_loss Res2Net/Res2Net-maskrcnn/maskrcnn_benchmark/layers/smooth_l1_loss.py community (archive-listed) ran MIT recorded; this copy not marked cleared · pointer only · e261fa29066b37e5 · report
interpolate Res2Net/Res2Net-maskrcnn/maskrcnn_benchmark/layers/misc.py community (archive-listed) unverified MIT recorded; this copy not marked cleared · pointer only · 2902bf4410253ff7 · report
res2net101_26w_4s Res2Net/Res2Net-PretrainedModels/res2net.py community (archive-listed) unverified no licence file found · pointer only · 3517e013785b7ac0 · report
res2net50 Res2Net/Res2Net-PretrainedModels/res2net.py community (archive-listed) unverified no licence file found · pointer only · a1acf33b04b5a6e5 · report
res2net50_26w_4s Res2Net/Res2Net-PretrainedModels/res2net.py community (archive-listed) unverified no licence file found · pointer only · f4eeaacba23fd852 · report
resnet50 ChrisMats/Res2Net/res2net.py community (archive-listed) unverified no licence file found · pointer only · 54f971e7a629c97e · report
sigmoid_focal_loss_cpu Res2Net/Res2Net-maskrcnn/maskrcnn_benchmark/layers/sigmoid_focal_loss.py community (archive-listed) unverified MIT recorded; this copy not marked cleared · pointer only · 426fd02a2beaf826 · report
conv1x1 identical code first harvested elsewhere ran · our draft was wrong licence of this copy not recorded · d9def42110729a85 · report

Tasks

Image ClassificationInstance SegmentationMedical Image ClassificationObject DetectionRGB Salient Object DetectionSalient Object Detectionobject-detection

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification CIFAR-100 Res2NeXt-29 Percentage correct 83.44 #88 of 211 Archive leaderboard report
Image Classification GasHisSDB Res2Net-50 Accuracy 98.68 #2 of 8 Archive leaderboard report
Image Classification GasHisSDB Res2Net-50 F1-Score 99.29 #2 of 8 Archive leaderboard report
Image Classification GasHisSDB Res2Net-50 Precision 99.91 #2 of 8 Archive leaderboard report
Image Classification ImageNet Res2Net-101 Top 1 Accuracy 81.23% #653 of 1060 Archive leaderboard report
Image Classification ImageNet Res2Net-50-299 Top 1 Accuracy 78.59% #824 of 1060 Archive leaderboard report
Instance Segmentation COCO minival Res2Net-101+HTC mask AP 41.3 #67 of 93 Archive leaderboard report
Instance Segmentation COCO minival Faster R-CNN (Res2Net-50) AP50 57.6 #89 of 93 Archive leaderboard report
Instance Segmentation COCO minival Faster R-CNN (Res2Net-50) APL 53.7 #89 of 93 Archive leaderboard report
Instance Segmentation COCO minival Faster R-CNN (Res2Net-50) APM 37.9 #89 of 93 Archive leaderboard report
Instance Segmentation COCO minival Faster R-CNN (Res2Net-50) APS 15.7 #89 of 93 Archive leaderboard report
Instance Segmentation COCO minival Faster R-CNN (Res2Net-50) mask AP 35.6 #89 of 93 Archive leaderboard report
Medical Image Classification NCT-CRC-HE-100K Res2Net-50 Accuracy (%) 93.37 #6 of 7 Archive leaderboard report
Medical Image Classification NCT-CRC-HE-100K Res2Net-50 F1-Score 96.25 #6 of 7 Archive leaderboard report
Medical Image Classification NCT-CRC-HE-100K Res2Net-50 Precision 99.93 #6 of 7 Archive leaderboard report
Medical Image Classification NCT-CRC-HE-100K Res2Net-50 Specificity 99.17 #6 of 7 Archive leaderboard report
Object Detection COCO minival Res2Net101+HTC AP50 66.5 #95 of 220 Archive leaderboard report
Object Detection COCO minival Res2Net101+HTC AP75 51.3 #95 of 220 Archive leaderboard report
Object Detection COCO minival Res2Net101+HTC APL 62.1 #95 of 220 Archive leaderboard report
Object Detection COCO minival Res2Net101+HTC APM 51.6 #95 of 220 Archive leaderboard report
Object Detection COCO minival Res2Net101+HTC APS 28.6 #95 of 220 Archive leaderboard report
Object Detection COCO minival Res2Net101+HTC box AP 47.5 #95 of 220 Archive leaderboard report
Object Detection COCO minival Faster R-CNN (Res2Net-50) AP50 53.6 #210 of 220 Archive leaderboard report
Object Detection COCO minival Faster R-CNN (Res2Net-50) APL 51.1 #210 of 220 Archive leaderboard report
Object Detection COCO minival Faster R-CNN (Res2Net-50) APM 38.3 #210 of 220 Archive leaderboard report
Object Detection COCO minival Faster R-CNN (Res2Net-50) APS 14 #210 of 220 Archive leaderboard report
Object Detection COCO minival Faster R-CNN (Res2Net-50) box AP 33.7 #210 of 220 Archive leaderboard report
RGB Salient Object Detection DUT-OMRON DSS (Res2Net-50) F-measure 0.800 #13 of 18 Archive leaderboard report
RGB Salient Object Detection DUT-OMRON DSS (Res2Net-50) MAE 0.071 #13 of 18 Archive leaderboard report
RGB Salient Object Detection ECSSD DSS (Res2Net-50) F-measure 0.926 #9 of 14 Archive leaderboard report
RGB Salient Object Detection ECSSD DSS (Res2Net-50) MAE 0.056 #9 of 14 Archive leaderboard report
RGB Salient Object Detection HKU-IS DSS (Res2Net-50) F-measure 0.905 #9 of 14 Archive leaderboard report
RGB Salient Object Detection HKU-IS DSS (Res2Net-50) MAE 0.05 #9 of 14 Archive leaderboard report
RGB Salient Object Detection PASCAL-S DSS (Res2Net-50) F-measure 0.841 #8 of 13 Archive leaderboard report
RGB Salient Object Detection PASCAL-S DSS (Res2Net-50) MAE 0.099 #8 of 13 Archive leaderboard report
Semantic Segmentation PASCAL VOC 2012 val Deeplab v3+ (Res2Net-101) mIoU 79.3% #12 of 29 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: Res2Net, Res2Net Block

1x1 ConvolutionAverage PoolingBatch NormalizationBottleneck Residual BlockConvolutionDLAFaster R-CNNGlobal Average PoolingGrouped ConvolutionKaiming InitializationMask R-CNNMax PoolingRPNRandom Horizontal FlipRandom Resized CropReLURes2NetRes2Net BlockResNeXtResNeXt BlockResidual BlockResidual ConnectionRoIAlignRoIPoolSGD with MomentumSoftmaxStep DecayWeight 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