Papers › YOLOX: Exceeding YOLO Series in 2021

YOLOX: Exceeding YOLO Series in 2021

18 Jul 2021arXiv:2107.08430archive 2025-07-28

Zheng Ge, Songtao Liu, Feng Wang, Zeming Li, Jian Sun

In this report, we present some experienced improvements to YOLO series, forming a new high-performance detector -- YOLOX. We switch the YOLO detector to an anchor-free manner and conduct other advanced detection techniques, i.e., a decoupled head and the leading label assignment strategy SimOTA to achieve state-of-the-art results across a large scale range of models: For YOLO-Nano with only 0.91M parameters and 1.08G FLOPs, we get 25.3% AP on COCO, surpassing NanoDet by 1.8% AP; for YOLOv3, one of the most widely used detectors in industry, we boost it to 47.3% AP on COCO, outperforming the current best practice by 3.0% AP; for YOLOX-L with roughly the same amount of parameters as YOLOv4-CSP, YOLOv5-L, we achieve 50.0% AP on COCO at a speed of 68.9 FPS on Tesla V100, exceeding YOLOv5-L by 1.8% AP. Further, we won the 1st Place on Streaming Perception Challenge (Workshop on Autonomous Driving at CVPR 2021) using a single YOLOX-L model. We hope this report can provide useful experience for developers and researchers in practical scenes, and we also provide deploy versions with ONNX, TensorRT, NCNN, and Openvino supported. Source code is at https://github.com/Megvii-BaseDetection/YOLOX.

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="2107.08430")

Code

Syntology Ran 1 of 23 code samples harvested from 7 repositories linked to this paper; 22 have no recorded run. Of those that ran: 1 ran · our draft was wrong.

By repository: community (archive-listed): 23 samples from 7 repositories, 1 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

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

Megvii-BaseDetection/YOLOX officialmentioned in papermentioned on GitHubpytorchApache-2.0 report
2023-MindSpore-1/ms-code-182 mentioned on GitHubmindspore report
2023-MindSpore-1/ms-code-31 mentioned on GitHubmindsporeApache-2.0 report
DataXujing/YOLOX- mentioned on GitHubpytorchApache-2.0 report
Deci-AI/super-gradients mentioned on GitHubpytorch report
MegEngine/YOLOX mentioned on GitHubpytorchApache-2.0 report
NEUAutonomousDriving408/CIVCAutonomousDriving mentioned on GitHubpytorchGPL-2.0 report
StephenStorm/YOLOX mentioned on GitHubpytorchApache-2.0 report
apolloauto/apollo-model-yolox mentioned on GitHubpytorchApache-2.0 report
chenyicai-0611/YOLOX-Flask-deployment mentioned on GitHubpytorchApache-2.0 report
jesse01/paddle-yolox mentioned on GitHubpaddle report
jinsheng124/yolox mentioned on GitHubpytorch report
kisna-aryan/YOLOX mentioned on GitHubpytorchApache-2.0 report
liuyuan000/yolox_sar mentioned on GitHubpytorchApache-2.0 report
middleprince/YOLOX-SS mentioned on GitHubpytorchApache-2.0 report
newsun-boki/yolovx-openvino-c--video-infer mentioned on GitHubpytorchApache-2.0 report
newsun-boki/yolox-openvino-video-infer mentioned on GitHubpytorchApache-2.0 report
open-edge-platform/geti mentioned on GitHubpytorchApache-2.0 report
pistachio0812/YOLOX mentioned on GitHubpytorchApache-2.0 report
pmj110119/yolox mentioned on GitHubpytorchApache-2.0 report
qy1994-0919/cfpnet mentioned on GitHubpytorchApache-2.0 report
texasinstruments/edgeai-yolox mentioned on GitHubpytorchNOASSERTION report
wangdongdut/dut-anti-uav mentioned on GitHubApache-2.0 report
xiyie/yolox mentioned on GitHubpytorchApache-2.0 report
zhangming8/yolox-pytorch mentioned on GitHubpytorch report
mszpc/yolox mindsporeApache-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

23 samples harvested; 1 ran; 0 honoured the contract we drafted; 22 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.

1ran · our draft was wrong
22unverified

Licence: 0 of the 23 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 7 repositories linked to this paper, official or community; each sample names its own and says which. “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.

merge 2023-MindSpore-1/ms-code-31/model_utils/config.py community (archive-listed) ran · our draft was wrong Apache-2.0 (permissive) · b9e24c3e3e0e88a7 · report
calculate_gain 2023-MindSpore-1/ms-code-31/src/initializer.py community (archive-listed) unverified Apache-2.0 (permissive) · e460868d0e106ef7 · report
create_yolox_model apolloauto/apollo-model-yolox/yolox/models/build.py community (archive-listed) unverified Apache-2.0 (permissive) · 7111f135eb8bfff5 · report
create_yolox_model kisna-aryan/YOLOX/yolox/models/build.py community (archive-listed) unverified Apache-2.0 (permissive) · 1c1d860f1cb790da · report
fuse_conv_and_bn DataXujing/YOLOX-/yolox/utils/model_utils.py community (archive-listed) unverified Apache-2.0 (permissive) · d3111a0ae5602114 · report
fuse_conv_and_bn apolloauto/apollo-model-yolox/yolox/utils/model_utils.py community (archive-listed) unverified Apache-2.0 (permissive) · 57bfa676482bc295 · report
fuse_model DataXujing/YOLOX-/yolox/utils/model_utils.py community (archive-listed) unverified Apache-2.0 (permissive) · bce7fbf2f1717430 · report
fuse_model apolloauto/apollo-model-yolox/yolox/utils/model_utils.py community (archive-listed) unverified Apache-2.0 (permissive) · 927238a3415f3aa2 · report
get_activation DataXujing/YOLOX-/yolox/models/network_blocks.py community (archive-listed) unverified Apache-2.0 (permissive) · 6e3d5f4c3ce305d7 · report
get_image_list chenyicai-0611/YOLOX-Flask-deployment/YOLOX-flask/backend/predict.py community (archive-listed) unverified Apache-2.0 (permissive) · 2f7db76440582d33 · report
get_logger 2023-MindSpore-1/ms-code-31/src/logger.py community (archive-listed) unverified Apache-2.0 (permissive) · bb879be65ea0ad0d · report
get_model_info apolloauto/apollo-model-yolox/yolox/utils/model_utils.py community (archive-listed) unverified Apache-2.0 (permissive) · 624c83e485703cd7 · report
kaiming_uniform_ 2023-MindSpore-1/ms-code-31/src/initializer.py community (archive-listed) unverified Apache-2.0 (permissive) · ac69eca3db22245e · report
parse_cli_to_yaml 2023-MindSpore-1/ms-code-31/model_utils/config.py community (archive-listed) unverified Apache-2.0 (permissive) · 738456e1b7388e7e · report
parse_yaml 2023-MindSpore-1/ms-code-31/model_utils/config.py community (archive-listed) unverified Apache-2.0 (permissive) · a5a844a8ffaaa722 · report
reduce_loss StephenStorm/YOLOX/yolox/models/VariFocalLoss.py community (archive-listed) unverified Apache-2.0 (permissive) · 93358ba7aa223343 · report
replace_module MegEngine/YOLOX/yolox/utils/model_utils.py community (archive-listed) unverified Apache-2.0 (permissive) · 14dbd04931eace27 · report
varifocal_loss StephenStorm/YOLOX/yolox/models/VariFocalLoss.py community (archive-listed) unverified Apache-2.0 (permissive) · 48e94cb9aba6d73d · report
weight_reduce_loss StephenStorm/YOLOX/yolox/models/VariFocalLoss.py community (archive-listed) unverified Apache-2.0 (permissive) · 8553862254d49840 · report
yolox_nano apolloauto/apollo-model-yolox/yolox/models/build.py community (archive-listed) unverified Apache-2.0 (permissive) · 706978d51729ca8d · report
yolox_nano kisna-aryan/YOLOX/yolox/models/build.py community (archive-listed) unverified Apache-2.0 (permissive) · 4657da89fcf077b4 · report
yolox_tiny apolloauto/apollo-model-yolox/yolox/models/build.py community (archive-listed) unverified Apache-2.0 (permissive) · af7e14a4cec878b8 · report
yolox_tiny kisna-aryan/YOLOX/yolox/models/build.py community (archive-listed) unverified Apache-2.0 (permissive) · 6bbd9ef4e16fa647 · report

Tasks

2D Object DetectionAutonomous DrivingObject DetectionReal-Time Object Detection

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
2D Object Detection CeyMo YOLOX mAP 57.7 #4 of 5 Archive leaderboard report
Object Detection COCO test-dev YOLOX-x(Modified CSP v5, 640x640, single-scale) box mAP 51.5 #78 of 225 Archive leaderboard report
Object Detection COCO test-dev YOLOX-X (Modified CSP v5) AP50 69.6 #83 of 225 Archive leaderboard report
Object Detection COCO test-dev YOLOX-X (Modified CSP v5) AP75 55.7 #83 of 225 Archive leaderboard report
Object Detection COCO test-dev YOLOX-X (Modified CSP v5) APL 66.1 #83 of 225 Archive leaderboard report
Object Detection COCO test-dev YOLOX-X (Modified CSP v5) APM 56.1 #83 of 225 Archive leaderboard report
Object Detection COCO test-dev YOLOX-X (Modified CSP v5) APS 31.2 #83 of 225 Archive leaderboard report
Object Detection COCO test-dev YOLOX-X (Modified CSP v5) Params (M) 99.1 #83 of 225 Archive leaderboard report
Object Detection COCO test-dev YOLOX-X (Modified CSP v5) box mAP 51.2 #83 of 225 Archive leaderboard report
Object Detection COCO test-dev YOLOX-Darknet53(Darknet53, 640x640, single-scale) box mAP 48.0 #111 of 225 Archive leaderboard report
Object Detection COCO-O YOLOX-X Average mAP 30.3 #19 of 45 Archive leaderboard report
Object Detection COCO-O YOLOX-X Effective Robustness 7.26 #19 of 45 Archive leaderboard report
Object Detection COCO-O YOLOX-S Average mAP 20.6 #29 of 45 Archive leaderboard report
Object Detection COCO-O YOLOX-S Effective Robustness 2.48 #29 of 45 Archive leaderboard report
Object Detection WaterScenes YOLOX-M mAP@50-95 57.8 #2 of 4 Archive leaderboard report
Real-Time Object Detection Argoverse-HD (Detection-Only, Test) YOLOX AP 41.1 #1 of 3 Archive leaderboard report
Real-Time Object Detection Argoverse-HD (Detection-Only, Val) YOLOX AP 47.42 #1 of 2 Archive leaderboard report
Real-Time Object Detection Argoverse-HD (Full-Stack, Test) YOLOX AP 41.1 #1 of 3 Archive leaderboard report
Real-Time Object Detection COCO (Common Objects in Context) YOLOv5-X FPS (V100, b=1) 62.5 #55 of 82 Archive leaderboard report
Real-Time Object Detection COCO (Common Objects in Context) YOLOv5-X box AP 50.4 #55 of 82 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

1x1 ConvolutionAverage PoolingBatch NormalizationCSPDarknet53ConvolutionGlobal Average PoolingLogistic RegressionResidual ConnectionSoftmaxYOLOYOLOXYOLOv3k-Means Clustering

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