Browse › Computer Vision › Object Detection › COCO minival
Object Detection archive 2025-07-28
COCO minival Benchmark (Object Detection)
Over time archive 2025-07-28
The chart needs JavaScript; the table below carries every value.
Direction inferred from the metric name, not from the archive: box AP (higher is better), Params (M) (lower is better). Not inferred (points only, no best-so-far line): AP50, AP75, APS, APM, APL. Points are placed at the row's paper date; 220 of 220 rows carry one.
Results archive 2025-07-28
Archive rows end at the archive snapshot, 2025-07-28: no result published after that date is in this table. Rank is the archive's row order at that snapshot; not re-ranked here. Metric values are the archive's strings. Column headers sort the table in your browser; each row keeps its archive rank.
| Paper | Code | Ran Syntology | Report | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | PE_spatial (DETA) | 66.0 | 1900 | ✓ | Paper | Code | 2025 | 12 of 18 ran · 6 unverified | report | |||||
| 2 | Co-DETR | 65.9 | 314 | ✓ | Paper | Code | 2022 | 0 of 5 ran · 5 unverified | report | |||||
| 3 | M3I Pre-training (InternImage-H) | 65.0 | ✓ | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 4 | InternImage-H | 65.0 | ✓ | Paper | Code | 2022 | 2 of 4 ran · 2 unverified | report | ||||||
| 5 | Co-DETR (Swin-L) | 64.7 | 218 | ✓ | Paper | Code | 2022 | 0 of 5 ran · 5 unverified | report | |||||
| 6 | Focal-Stable-DINO (Focal-Huge, no TTA) | 64.6 | 81.5 | 71.4 | 50.4 | 68.5 | 78.5 | 689 | ✓ | Paper | Code | 2023 | linked, not harvested | report |
| 7 | EVA | 64.5 | 82.1 | 70.8 | 49.4 | 68.4 | 78.5 | ✓ | Paper | Code | 2022 | 1 of 3 ran · 2 unverified | report | |
| 8 | ViT-CoMer | 64.3 | 363 | – | Paper | Code | 2024 | linked, not harvested | report | |||||
| 9 | FocalNet-H (DINO) | 64.2 | ✓ | Paper | Code | 2022 | 0 of 2 ran · 2 unverified | report | ||||||
| 10 | InternImage-XL | 64.2 | ✓ | Paper | Code | 2022 | 2 of 4 ran · 2 unverified | report | ||||||
| 11 | CP-DETR-L Swin-L(Fine tuning separately in COCO) | 64.1 | ✓ | Paper | – | 2024 | no code linked | report | ||||||
| 12 | RevCol-H(DINO) | 63.8 | ✓ | Paper | Code | 2022 | 8 of 11 ran · 3 unverified | report | ||||||
| 13 | DINO (Swin-L) | 63.2 | – | Paper | Code | 2022 | 7 of 15 ran · 8 unverified | report | ||||||
| 14 | Grounding DINO | 63.0 | ✓ | Paper | Code | 2023 | 2 of 5 ran · 3 unverified | report | ||||||
| 15 | SwinV2-G (HTC++) | 62.5 | ✓ | Paper | Code | 2021 | 3 of 30 ran · 27 unverified | report | ||||||
| 16 | Florence-CoSwin-H | 62 | ✓ | Paper | Code | 2021 | linked, not harvested | report | ||||||
| 17 | GLEE-Pro | 62.0 | ✓ | Paper | Code | 2023 | 8 of 13 ran · 5 unverified | report | ||||||
| 18 | ViTDet, ViT-H Cascade (multiscale) | 61.3 | – | Paper | Code | 2022 | 1 of 4 ran · 3 unverified | report | ||||||
| 19 | GLIP (Swin-L, multi-scale) | 60.8 | ✓ | Paper | Code | 2021 | 2 of 2 ran · 0 unverified | report | ||||||
| 20 | Soft Teacher + Swin-L (HTC++, multi-scale) | 60.7 | ✓ | Paper | Code | 2021 | linked, not harvested | report | ||||||
| 21 | UNINEXT-H | 60.6 | 77.5 | 66.7 | 45.1 | 64.8 | 75.3 | ✓ | Paper | Code | 2023 | 3 of 4 ran · 1 unverified | report | |
| 22 | ViT-Adapter-L (HTC++, BEiTv2 pretrain, multi-scale) | 60.5 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 23 | ViTDet, ViT-H Cascade | 60.4 | – | Paper | Code | 2022 | 1 of 4 ran · 3 unverified | report | ||||||
| 24 | GLEE-Plus | 60.4 | ✓ | Paper | Code | 2023 | 8 of 13 ran · 5 unverified | report | ||||||
| 25 | DyHead (Swin-L, multi scale, self-training) | 60.3 | 78.2 | 74.2 | ✓ | Paper | Code | 2021 | 7 of 9 ran · 2 unverified | report | ||||
| 26 | ViT-Adapter-L (HTC++, BEiT pretrain, multi-scale) | 60.2 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 27 | Soft Teacher+Swin-L(HTC++, single scale) | 60.1 | ✓ | Paper | Code | 2021 | linked, not harvested | report | ||||||
| 28 | CBNetV2 (Dual-Swin-L HTC, multi-scale) | 59.6 | – | Paper | Code | 2021 | 2 of 3 ran · 1 unverified | report | ||||||
| 29 | Frozen Backbone, SwinV2-G-ext22K (HTC) | 59.3 | – | Paper | – | 2022 | no code linked | report | ||||||
| 30 | HorNet-L | 59.2 | – | Paper | Code | 2022 | 1 of 1 ran · 0 unverified | report | ||||||
| 31 | MOAT-3 (IN-22K pretraining, single-scale) | 59.2 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 32 | CBNetV2 (Dual-Swin-L HTC, multi-scale) | 59.1 | – | Paper | Code | 2021 | 2 of 3 ran · 1 unverified | report | ||||||
| 33 | Focal-L (DyHead, multi-scale) | 58.7 | 77.2 | 73.4 | – | Paper | Code | 2021 | 3 of 3 ran · 0 unverified | report | ||||
| 34 | MViTv2-L (Cascade Mask R-CNN, multi-scale, IN21k pre-train) | 58.7 | – | Paper | Code | 2021 | linked, not harvested | report | ||||||
| 35 | MOAT-2 (IN-22K pretraining, single-scale) | 58.5 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 36 | DyHead (Swin-L, multi scale) | 58.4 | 76.8 | 44.5 | 62.2 | 73.2 | – | Paper | Code | 2021 | 7 of 9 ran · 2 unverified | report | ||
| 37 | Swin-L (HTC++, multi scale) | 58 | – | Paper | Code | 2021 | 108 of 207 ran · 99 unverified | report | ||||||
| 38 | MOAT-1 (IN-1K pretraining, single-scale) | 57.7 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 39 | UM-MAE(HTC++, Swin-L, IN1K) | 57.4 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 40 | YOLOv6-L6(46 fps, 1280, V100) | 57.2 | 74.5 | – | Paper | Code | 2023 | 2 of 2 ran · 0 unverified | report | |||||
| 41 | Swin-L (HTC++, single scale) | 57.1 | – | Paper | Code | 2021 | 108 of 207 ran · 99 unverified | report | ||||||
| 42 | TransNeXt-Base (IN-1K pretrain, DINO 1x) | 57.1 | – | Paper | Code | 2023 | 5 of 5 ran · 0 unverified | report | ||||||
| 43 | Cascade Eff-B7 NAS-FPN (1280, self-training Copy Paste, single-scale) | 57.0 | ✓ | Paper | Code | 2020 | 2 of 2 ran · 0 unverified | report | ||||||
| 44 | TransNeXt-Small (IN-1K pretrain, DINO 1x) | 56.6 | – | Paper | Code | 2023 | 5 of 5 ran · 0 unverified | report | ||||||
| 45 | QueryInst (single scale) | 56.1 | 75.8 | 61.7 | 40.2 | 59.8 | 71.5 | – | Paper | Code | 2021 | 1 of 3 ran · 2 unverified | report | |
| 46 | MViTv2-H (Cascade Mask R-CNN, single-scale, IN21k pre-train) | 56.1 | – | Paper | Code | 2021 | linked, not harvested | report | ||||||
| 47 | MOAT-0 (IN-1K pretraining, single-scale) | 55.9 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 48 | TransNeXt-Tiny (IN-1K pretrain, DINO 1x) | 55.7 | – | Paper | Code | 2023 | 5 of 5 ran · 0 unverified | report | ||||||
| 49 | YOLOv4-P7 CSP-P7 (single-scale, 16 fps) | 55.4 | 73.3 | 60.7 | 38.1 | 59.5 | 67.4 | – | Paper | Code | 2020 | 1 of 1 ran · 0 unverified | report | |
| 50 | tiny-MOAT-3 (IN-1K pretraining, single-scale) | 55.2 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 51 | FAN-L-Hybrid | 55.1 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 52 | Hiera-L | 55 | – | Paper | Code | 2023 | 0 of 6 ran · 6 unverified | report | ||||||
| 53 | GLEE-Lite | 55.0 | ✓ | Paper | Code | 2023 | 8 of 13 ran · 5 unverified | report | ||||||
| 54 | TEC(VIT-B, Mask-RCNN) | 54.6 | – | Paper | Code | 2022 | 6 of 8 ran · 2 unverified | report | ||||||
| 55 | Cascade Eff-B7 NAS-FPN (1280) | 54.5 | – | Paper | Code | 2020 | 2 of 2 ran · 0 unverified | report | ||||||
| 56 | CAE (ViT-L, Mask R-CNN, 1x schedule) | 54.5 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 57 | MViTv2-L (Cascade Mask R-CNN, single-scale) | 54.3 | – | Paper | Code | 2021 | linked, not harvested | report | ||||||
| 58 | SpineNet-190 (1280, with Self-training on OpenImages, single-scale) | 54.2 | ✓ | Paper | Code | 2020 | linked, not harvested | report | ||||||
| 59 | Cascade RCNN-RS (SpineNet-143L, single scale) | 53.6 | 34.5 | 56.7 | 70.6 | – | Paper | Code | 2021 | linked, not harvested | report | |||
| 60 | UniverseNet-20.08d (Res2Net-101, DCN, multi-scale) | 53.5 | 70.8 | 58.9 | 36.9 | 57.5 | 68.1 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 61 | MAE (ViT-L, Mask R-CNN) | 53.3 | – | Paper | Code | 2021 | 71 of 137 ran · 66 unverified | report | ||||||
| 62 | Cascade RCNN-RS (ResNet-200, single scale) | 53.1 | 33.9 | 56.2 | 70.3 | – | Paper | Code | 2021 | linked, not harvested | report | |||
| 63 | tiny-MOAT-2 (IN-1K pretraining, single-scale) | 53.0 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 64 | MViT-L (Mask R-CNN, single-scale, IN21k pre-train) | 52.7 | – | Paper | Code | 2021 | linked, not harvested | report | ||||||
| 65 | ResNeSt-200 (multi-scale) | 52.47 | 71.00 | 57.07 | 36.80 | 56.36 | 66.29 | – | Paper | Code | 2020 | 8 of 48 ran · 40 unverified | report | |
| 66 | ActiveMLP-B (Cascade Mask R-CNN) | 52.3 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 67 | RetinaNet (SpineNet-190, 1536x1536) | 52.2 | – | Paper | Code | 2019 | linked, not harvested | report | ||||||
| 68 | EfficientDet-D7 (1536) | 52.1 | – | Paper | Code | 2019 | 11 of 70 ran · 59 unverified | report | ||||||
| 69 | tiny-MOAT-1 (IN-1K pretraining, single-scale) | 51.9 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 70 | GCNet (ResNeXt-101 + DCN + cascade + GC r4) | 51.8 | 70.4 | 56.1 | – | Paper | Code | 2020 | linked, not harvested | report | ||||
| 71 | ELSA-S (Cascade Mask RCNN) | 51.6 | 70.5 | 56.0 | – | Paper | Code | 2021 | 1 of 1 ran · 0 unverified | report | ||||
| 72 | FocalNet-T (LRF, Cascade Mask R-CNN) | 51.5 | 70.3 | 56.0 | – | Paper | Code | 2022 | 0 of 2 ran · 2 unverified | report | ||||
| 73 | DINO-5scale (24 epoch) | 51.3 | 69.1 | 56 | 34.5 | 54.2 | 65.8 | – | Paper | Code | 2022 | 7 of 15 ran · 8 unverified | report | |
| 74 | DINO-5scale (36 epoch) | 51.2 | 69 | 55.8 | 35 | 54.3 | 65.3 | – | Paper | Code | 2022 | 7 of 15 ran · 8 unverified | report | |
| 75 | ResNeSt-200-DCN (single-scale) | 50.91 | 69.53 | 55.40 | 32.67 | 54.66 | 65.83 | – | Paper | Code | 2020 | 8 of 48 ran · 40 unverified | report | |
| 76 | UniverseNet-20.08d (Res2Net-101, DCN, single-scale) | 50.9 | 69.5 | 55.4 | 33.5 | 55.5 | 65.8 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 77 | ResNeSt-200 (single-scale) | 50.54 | 68.78 | 55.17 | 54.2 | 63.9 | – | Paper | Code | 2020 | 8 of 48 ran · 40 unverified | report | ||
| 78 | tiny-MOAT-0 (IN-1K pretraining, single-scale) | 50.5 | – | Paper | Code | 2022 | linked, not harvested | report | ||||||
| 79 | MAE (ViT-B, Mask R-CNN) | 50.3 | – | Paper | Code | 2021 | 71 of 137 ran · 66 unverified | report | ||||||
| 80 | Sparse R-CNN (PVTv2-B2) | 50.1 | 69.5 | 54.9 | – | Paper | Code | 2021 | linked, not harvested | report | ||||
| 81 | Pix2seq (ViT-L) | 50.0 | ✓ | Paper | Code | 2021 | 10 of 12 ran · 2 unverified | report | ||||||
| 82 | DaViT-T (Mask R-CNN, 36 epochs) | 49.9 | – | Paper | Code | 2022 | 8 of 15 ran · 7 unverified | report | ||||||
| 83 | BoTNet 200 (Mask R-CNN, single scale, 72 epochs) | 49.7 | 71.3 | 54.6 | – | Paper | Code | 2021 | 26 of 49 ran · 23 unverified | report | ||||
| 84 | BoTNet 152 (Mask R-CNN, single scale, 72 epochs) | 49.5 | 71 | 54.2 | – | Paper | Code | 2021 | 26 of 49 ran · 23 unverified | report | ||||
| 85 | DN-Deformable-DETR-R50++ | 49.5 | 67.6 | 53.8 | 31.3 | 52.6 | 65.4 | 47 | – | Paper | Code | 2022 | 15 of 22 ran · 7 unverified | report |
| 86 | REGO-Deformable DETR-X101 | 49.1 | 67.5 | 53.1 | 30 | 52.6 | 65 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 87 | CenterMask+VoVNet99 (multi-scale) | 48.6 | 67.8 | – | Paper | Code | 2019 | 1 of 2 ran · 1 unverified | report | |||||
| 88 | Mask R-CNN (ResNeXt-152-FPN, cascade) | 48.6 | 66.8 | 52.9 | – | Paper | Code | 2018 | linked, not harvested | report | ||||
| 89 | UniverseNet-20.08 (Res2Net-50, DCN, single-scale) | 48.5 | 67.0 | 52.6 | 30.6 | 52.7 | 62.7 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 90 | XCiT-M24/8 | 48.5 | – | Paper | Code | 2021 | 3 of 14 ran · 11 unverified | report | ||||||
| 91 | ELSA-S (Mask RCNN) | 48.3 | 70.4 | 52.9 | – | Paper | Code | 2021 | 1 of 1 ran · 0 unverified | report | ||||
| 92 | XCiT-S24/8 | 48.1 | – | Paper | Code | 2021 | 3 of 14 ran · 11 unverified | report | ||||||
| 93 | GCNet (ResNeXt-101 + DCN + cascade + GC r16) | 47.9 | 66.9 | 52.2 | – | Paper | Code | 2019 | linked, not harvested | report | ||||
| 94 | MAE-Det(MAE-Det-L+GFLV2) | 47.8 | 65.5 | 52.2 | 30.3 | 51.9 | 61.1 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 95 | Res2Net101+HTC | 47.5 | 66.5 | 51.3 | 28.6 | 51.6 | 62.1 | – | Paper | Code | 2019 | 3 of 9 ran · 6 unverified | report | |
| 96 | Mask R-CNN (ResNet-101-FPN, GN, Cascade) | 47.4 | – | Paper | Code | 2018 | linked, not harvested | report | ||||||
| 97 | Pix2seq (R50-C4) | 47.3 | – | Paper | Code | 2021 | 10 of 12 ran · 2 unverified | report | ||||||
| 98 | Pix2seq (ViT-B) | 47.1 | – | Paper | Code | 2021 | 10 of 12 ran · 2 unverified | report | ||||||
| 99 | HTC (HRNetV2p-W48) | 47.0 | 28.8 | 50.3 | 62.2 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | |||
| 100 | PatchConvNet-S120 (Mask R-CNN) | 47.0 | – | Paper | Code | 2021 | 1 of 2 ran · 1 unverified | report | ||||||
| 101 | RPDet (ResNeXt-101-DCN, multi-scale) | 46.8 | – | Paper | Code | 2019 | 1 of 3 ran · 2 unverified | report | ||||||
| 102 | DAB-DETR-DC5-R101 | 46.6 | 67 | 50.2 | 28.1 | 50.5 | 64.1 | 63 | – | Paper | Code | 2022 | 6 of 11 ran · 5 unverified | report |
| 103 | DyHead (ResNet-101) | 46.5 | – | Paper | Code | 2021 | 7 of 9 ran · 2 unverified | report | ||||||
| 104 | Mask R-CNN (ResNeXt-152-FPN) | 46.4 | 67.1 | 51.1 | – | Paper | Code | 2018 | linked, not harvested | report | ||||
| 105 | RPDet (ResNet-101-DCN, multi-scale) | 46.4 | – | Paper | Code | 2019 | 1 of 3 ran · 2 unverified | report | ||||||
| 106 | PatchConvNet-S60 (Mask R-CNN) | 46.4 | – | Paper | Code | 2021 | 1 of 2 ran · 1 unverified | report | ||||||
| 107 | Cascade Mask R-CNN (ResNet-50) | 46.3 | 64.3 | 50.5 | – | Paper | Code | 2015 | 230 of 377 ran · 147 unverified | report | ||||
| 108 | HoughNet (HG-104, MS) | 46.1 | 64.6 | 50.3 | 30.0 | 48.8 | 59.7 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 109 | Mask R-CNN (HRNetV2p-W48, cascade) | 46.0 | 27.5 | 60.1 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | ||||
| 110 | Conditional DETR-DC5-R101 | 45.9 | 66.8 | 49.5 | 27.2 | 50.3 | 63.3 | 63 | – | Paper | Code | 2021 | 5 of 9 ran · 4 unverified | report |
| 111 | BoTNet 50 (72 epochs) | 45.9 | – | Paper | Code | 2021 | 26 of 49 ran · 23 unverified | report | ||||||
| 112 | Sparse R-CNN (ResNet-101, learnable proposals, random crop aug, FPN) | 45.6 | 64.6 | 49.5 | 28.3 | 48.3 | 61.6 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 113 | CenterMask+VoVNetV2-99 (single-scale) | 45.6 | 29.2 | 58.8 | – | Paper | Code | 2019 | 1 of 2 ran · 1 unverified | report | ||||
| 114 | HTC (HRNetV2p-W32) | 45.3 | 27.0 | 48.4 | 59.5 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | |||
| 115 | Anchor DETR-DC5-R101 | 45.1 | 65.7 | 48.8 | 25.8 | 49.4 | 61.6 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 116 | Conditional DETR-DC5-R50 | 45.1 | 65.4 | 48.5 | 25.3 | 49 | 62.2 | 44 | – | Paper | Code | 2021 | 5 of 9 ran · 4 unverified | report |
| 117 | Mask R-CNN (ResNeXt-152 + 1 NL) | 45.0 | 67.8 | 48.9 | – | Paper | Code | 2017 | 3 of 4 ran · 1 unverified | report | ||||
| 118 | Pix2seq (R101-DC5) | 45.0 | 63.2 | 48.6 | 28.2 | 48.9 | 60.4 | – | Paper | Code | 2021 | 10 of 12 ran · 2 unverified | report | |
| 119 | Mask R-CNN-FPN (AOGNet-40M) | 44.9 | 66.2 | 49.1 | – | Paper | Code | 2019 | linked, not harvested | report | ||||
| 120 | DETR-DC5 (ResNet-101) | 44.9 | 64.7 | 47.7 | 23.7 | 49.5 | 62.3 | – | Paper | Code | 2020 | 59 of 92 ran · 33 unverified | report | |
| 121 | Mask R-CNN (VoVNetV2-99, single-scale) | 44.9 | 28.5 | 57.7 | – | Paper | Code | 2019 | 1 of 2 ran · 1 unverified | report | ||||
| 122 | R3-CNN (ResNet-50-FPN, DCN) | 44.8 | 64.3 | 48.9 | 26.6 | 48.3 | 59.6 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 123 | RPDet (ResNet-101-DCN, multi-scale train) | 44.8 | – | Paper | Code | 2019 | 1 of 3 ran · 2 unverified | report | ||||||
| 124 | RetinaNet (ViL-Base, multi-scale, 3x) | 44.7 | 47.6 | 29.9 | 48 | 58.1 | – | Paper | Code | 2021 | 6 of 10 ran · 4 unverified | report | ||
| 125 | Cascade R-CNN (HRNetV2p-W48) | 44.6 | 62.7 | 48.7 | 26.3 | 48.1 | 58.5 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | |
| 126 | CenterMask+VoVNetV2-57 (single-scale) | 44.6 | 27.7 | 48.3 | – | Paper | Code | 2019 | 1 of 2 ran · 1 unverified | report | ||||
| 127 | Conditional DETR-R101 | 44.5 | 65.6 | 47.5 | 23.6 | 48.4 | 63.6 | 63 | – | Paper | Code | 2021 | 5 of 9 ran · 4 unverified | report |
| 128 | Sparse R-CNN (ResNet-50, learnable proposals, random crop aug, FPN) | 44.5 | 63.4 | 48.2 | 26.9 | 47.2 | 59.5 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 129 | GFL (ResNet-50) | 44.5 | 63.0 | 48.3 | – | Paper | Code | 2015 | 230 of 377 ran · 147 unverified | report | ||||
| 130 | RPDet (ResNeXt-101-DCN) | 44.5 | – | Paper | Code | 2019 | 1 of 3 ran · 2 unverified | report | ||||||
| 131 | CenterMask+X101-32x8d (single-scale) | 44.4 | 26.7 | 57.1 | – | Paper | Code | 2019 | 1 of 2 ran · 1 unverified | report | ||||
| 132 | RetinaNet (ViL-Base) | 44.3 | 65.5 | 47.1 | 28.9 | 47.9 | 58.3 | – | Paper | Code | 2021 | 6 of 10 ran · 4 unverified | report | |
| 133 | R3-CNN (ResNet-50-FPN, GC-Net) | 44.3 | 64.1 | 48.4 | 27 | 47.1 | 58.9 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 134 | Anchor DETR-DC5-R50 | 44.2 | 64.7 | 47.5 | 24.7 | 48.2 | 60.6 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 135 | DAB-DETR-R101 | 44.1 | 64.7 | 47.2 | 24.1 | 48.2 | 62.9 | 63 | – | Paper | Code | 2022 | 6 of 11 ran · 5 unverified | report |
| 136 | Faster RCNN-R101-FPN+ | 44 | 63.9 | 47.8 | 27.2 | 48.1 | 56 | – | Paper | Code | 2020 | 59 of 92 ran · 33 unverified | report | |
| 137 | Cascade R-CNN (HRNetV2p-W32) | 43.7 | 61.7 | 47.7 | 25.6 | 46.5 | 57.4 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | |
| 138 | Sparse R-CNN (ResNet-101, FPN) | 43.5 | 62.1 | 47.2 | 26.1 | 46.3 | 59.7 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 139 | ATSS (ResNet-50) | 43.5 | 61.9 | 47.0 | – | Paper | Code | 2015 | 230 of 377 ran · 147 unverified | report | ||||
| 140 | PVT-Large (RetinaNet 3x,MS) | 43.4 | 63.6 | 46.1 | 26.1 | 46.0 | 59.5 | – | Paper | Code | 2021 | 22 of 30 ran · 8 unverified | report | |
| 141 | ExtremeNet (Hourglass-104, multi-scale) | 43.3 | 59.6 | 46.8 | 25.7 | 46.6 | 59.4 | – | Paper | Code | 2019 | 1 of 5 ran · 4 unverified | report | |
| 142 | Pix2seq (R50-DC5 ) | 43.2 | 61.0 | 46.1 | 26.6 | 47 | 58.6 | – | Paper | Code | 2021 | 10 of 12 ran · 2 unverified | report | |
| 143 | HTC (cascade) | 43.2 | 59.4 | 40.7 | 20.3 | 40.9 | 52.3 | – | Paper | Code | 2019 | 0 of 11 ran · 11 unverified | report | |
| 144 | Mask R-CNN-FPN (ResNeXt-101, GN+WS) | 43.12 | 64.15 | 47.11 | 25.49 | 47.19 | 56.39 | – | Paper | Code | 2019 | linked, not harvested | report | |
| 145 | HTC (HRNetV2p-W18) | 43.1 | 26.6 | 46.0 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | ||||
| 146 | Mask R-CNN (ResNet-101, DCNv2) | 43.1 | – | Paper | Code | 2018 | 2 of 13 ran · 11 unverified | report | ||||||
| 147 | Conditional DETR-R50 | 43 | 64 | 45.7 | 22.7 | 46.7 | 61.5 | 44 | – | Paper | Code | 2021 | 5 of 9 ran · 4 unverified | report |
| 148 | HoughNet (HG-104) | 43.0 | 62.2 | 46.9 | 25.5 | 47.6 | 55.8 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 149 | Faster R-CNN (FPN, X-volution) | 42.8 | 64 | 46.4 | 26.9 | 46 | 55 | – | Paper | – | 2021 | no code linked | report | |
| 150 | Cascade R-CNN (ResNet-101-FPN+, cascade) | 42.7 | 61.6 | 46.6 | 23.8 | 46.2 | 57.4 | – | Paper | Code | 2017 | 2 of 2 ran · 0 unverified | report | |
| 151 | PVT-Large (RetinaNet 1x) | 42.6 | 63.7 | 45.4 | 25.8 | 46.0 | 58.4 | – | Paper | Code | 2021 | 22 of 30 ran · 8 unverified | report | |
| 152 | CornerNet-Saccade (Hourglass-54) | 42.6 | 25.5 | 44.3 | 58.4 | – | Paper | Code | 2019 | 1 of 27 ran · 26 unverified | report | |||
| 153 | Pix2seq (R50) | 42.6 | – | Paper | Code | 2021 | 10 of 12 ran · 2 unverified | report | ||||||
| 154 | Mask R-CNN (ResNet-101-FPN, GroupNorm, long) | 42.3 | 62.8 | 46.2 | – | Paper | Code | 2018 | 5 of 15 ran · 10 unverified | report | ||||
| 155 | Sparse R-CNN (ResNet-50, FPN) | 42.3 | 61.2 | 45.7 | 26.7 | 44.6 | 57.6 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 156 | Mask R-CNN (HRNetV2p-W32) | 42.3 | 25.0 | 45.4 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | ||||
| 157 | DETR-ResNet50 with iRPE-K (300 epochs) | 42.3 | – | Paper | Code | 2021 | 7 of 9 ran · 2 unverified | report | ||||||
| 158 | TridentNet (ResNet-101) | 42 | 63.5 | 45.5 | 24.9 | 47 | 56.9 | – | Paper | Code | 2019 | linked, not harvested | report | |
| 159 | R3-CNN (ResNet-50-FPN) | 42 | 61 | 46.3 | 24.5 | 45.2 | 55.7 | – | Paper | Code | 2021 | linked, not harvested | report | |
| 160 | Faster R-CNN (HRNetV2p-W48) | 41.8 | 62.8 | 45.9 | 44.7 | 54.6 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | ||
| 161 | Faster R-CNN (LIP-ResNet-101) | 41.7 | 63.6 | 45.6 | 25.2 | 45.8 | – | Paper | Code | 2019 | linked, not harvested | report | ||
| 162 | Faster R-CNN (ResNet-101, DCNv2) | 41.7 | 22.2 | 45.8 | 58.7 | – | Paper | Code | 2018 | 2 of 13 ran · 11 unverified | report | |||
| 163 | FSAF (ResNeXt-101, anchor-based branches) | 41.6 | 62.4 | – | Paper | Code | 2019 | 0 of 8 ran · 8 unverified | report | |||||
| 164 | CornerNet-Saccade (Hourglass-104) | 41.4 | 23.8 | 43.5 | 57.1 | – | Paper | Code | 2019 | 1 of 27 ran · 26 unverified | report | |||
| 165 | Grid R-CNN (ResNet-101-FPN) | 41.3 | 60.3 | 44.4 | 23.4 | 45.8 | 54.1 | – | Paper | Code | 2018 | 0 of 3 ran · 3 unverified | report | |
| 166 | Cascade R-CNN (HRNetV2p-W18) | 41.3 | 59.2 | 44.9 | 23.7 | 44.2 | 54.1 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | |
| 167 | CenterNet511 (Hourglass-52) | 41.3 | 59.2 | 43.9 | 23.6 | 43.8 | 55.8 | – | Paper | Code | 2019 | 2 of 11 ran · 9 unverified | report | |
| 168 | RetinaMask (ResNet-101-FPN) | 41.1 | 60.2 | 44.1 | – | Paper | Code | 2019 | linked, not harvested | report | ||||
| 169 | PoolFormer-S36 (Mask R-CNN) | 41.0 | 63.1 | 44.8 | – | Paper | Code | 2021 | 2 of 3 ran · 1 unverified | report | ||||
| 170 | Faster R-CNN (HRNetV2p-W32) | 40.9 | 61.8 | 44.8 | 24.4 | 43.7 | 53.3 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | |
| 171 | VirTex Mask R-CNN (ResNet-50-FPN) | 40.9 | – | Paper | Code | 2020 | 0 of 1 ran · 1 unverified | report | ||||||
| 172 | Mask R-CNN (ResNet-101 + 1 NL) | 40.8 | 63.1 | 44.5 | – | Paper | Code | 2017 | 3 of 4 ran · 1 unverified | report | ||||
| 173 | Mask R-CNN (ResNet-50-FPN, GroupNorm, long) | 40.8 | 61.6 | 44.4 | – | Paper | Code | 2018 | 5 of 15 ran · 10 unverified | report | ||||
| 174 | RPDet (ResNet-50, multi-scale train) | 40.8 | – | Paper | Code | 2019 | 1 of 3 ran · 2 unverified | report | ||||||
| 175 | DETR-ResNet50 with iRPE-K (150 epochs) | 40.8 | – | Paper | Code | 2021 | 7 of 9 ran · 2 unverified | report | ||||||
| 176 | Faster R-CNN+aLRP Loss (ResNet-50, 500 scale) | 40.7 | 60.7 | 43.3 | – | Paper | Code | 2020 | 2 of 2 ran · 0 unverified | report | ||||
| 177 | PPDet (ResNet-101-FPN) | 40.5 | 59.5 | 44.2 | 25.4 | 44.7 | 52.3 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 178 | GCnet (ResNet-50-FPN, GRoIE) | 40.3 | 62.4 | 44 | 24.2 | 44.4 | 52.5 | – | Paper | Code | 2019 | linked, not harvested | report | |
| 179 | Mask R-CNN (ResNet-50-FPN, GroupNorm) | 40.3 | 61 | 44 | – | Paper | Code | 2018 | 5 of 15 ran · 10 unverified | report | ||||
| 180 | Cascade R-CNN (ResNet-50-FPN+) | 40.3 | 59.4 | 43.7 | 22.9 | 43.7 | 54.1 | – | Paper | Code | 2017 | 2 of 2 ran · 0 unverified | report | |
| 181 | ExtremeNet (Hourglass-104, single-scale) | 40.3 | 55.1 | 43.7 | 21.6 | 44.0 | 56.1 | – | Paper | Code | 2019 | 1 of 5 ran · 4 unverified | report | |
| 182 | RPDet (ResNet-101) | 40.3 | – | Paper | Code | 2019 | 1 of 3 ran · 2 unverified | report | ||||||
| 183 | RetinaNet+aLRP Loss (ResNet-50, 500 scale) | 40.2 | 60.3 | 42.3 | – | Paper | Code | 2020 | 2 of 2 ran · 0 unverified | report | ||||
| 184 | Mask R-CNN (ResNet-101-FPN) | 40.0 | – | Paper | Code | 2017 | 42 of 140 ran · 98 unverified | report | ||||||
| 185 | FPN+ | 39.8 | 61.3 | 43.3 | 22.9 | 43.3 | 52.6 | – | Paper | Code | 2016 | 16 of 51 ran · 35 unverified | report | |
| 186 | FoveaBox+aLRP Loss (ResNet-50, 500 scale) | 39.7 | 58.8 | 41.5 | – | Paper | Code | 2020 | 2 of 2 ran · 0 unverified | report | ||||
| 187 | Grid R-CNN (ResNet-50-FPN) | 39.6 | 58.3 | 42.4 | 22.6 | 43.8 | 51.5 | – | Paper | Code | 2018 | 0 of 3 ran · 3 unverified | report | |
| 188 | Mask R-CNN (ResNet-50, ACNet) | 39.5 | – | Paper | Code | 2019 | linked, not harvested | report | ||||||
| 189 | FSAF (ResNet-101, anchor-based branches) | 39.3 | 59.2 | – | Paper | Code | 2019 | 0 of 8 ran · 8 unverified | report | |||||
| 190 | Mask R-CNN (HRNetV2p-W18) | 39.2 | 41.7 | 51.0 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | ||||
| 191 | Mask R-CNN (ResNet-50 + 1 NL) | 39.0 | 61.1 | 41.9 | – | Paper | Code | 2017 | 3 of 4 ran · 1 unverified | report | ||||
| 192 | FoveaBox (ResNet-101-FPN, 800x800) | 38.9 | 58.4 | 41.5 | 22.3 | 43.5 | 51.7 | – | Paper | Code | 2019 | linked, not harvested | report | |
| 193 | FCOS (ResNet-50-FPN + improvements) | 38.6 | 57.4 | 41.4 | 22.3 | 42.5 | 49.8 | – | Paper | Code | 2019 | 13 of 40 ran · 27 unverified | report | |
| 194 | RPDet (ResNet-50) | 38.6 | – | Paper | Code | 2019 | 1 of 3 ran · 2 unverified | report | ||||||
| 195 | Libra R-CNN (ResNet-50 FPN) | 38.5 | 59.3 | 42.0 | 22.9 | 42.1 | 50.5 | – | Paper | Code | 2019 | linked, not harvested | report | |
| 196 | Mask R-CNN (ResNet-50-FPN, GRoIE) | 38.4 | 59.9 | 41.7 | 22.9 | 42.1 | 49.7 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 197 | CornerNet511 (Hourglass-104) | 38.4 | 53.8 | 40.9 | 18.6 | 40.5 | 51.8 | – | Paper | Code | 2018 | 3 of 11 ran · 8 unverified | report | |
| 198 | FoveaBox+Retina (ResNet-50) | 38.1 | 57.8 | 40.5 | – | Paper | Code | 2019 | linked, not harvested | report | ||||
| 199 | Faster R-CNN (HRNetV2p-W18) | 38.0 | 58.9 | 41.5 | 22.6 | 40.8 | 49.6 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | |
| 200 | FoveaBox (ResNet-101-FPN, 600x600) | 38 | 57.8 | 40.2 | 19.5 | 42.2 | 52.7 | – | Paper | Code | 2019 | linked, not harvested | report | |
| 201 | FSAF (ResNet-101) | 37.9 | 58.0 | – | Paper | Code | 2019 | 0 of 8 ran · 8 unverified | report | |||||
| 202 | Mask R-CNN (ResNet-50-FPN) | 37.7 | – | Paper | Code | 2017 | 42 of 140 ran · 98 unverified | report | ||||||
| 203 | Faster R-CNN (ResNet-50-FPN, GRoIE) | 37.5 | 59.2 | 40.6 | 22.3 | 41.5 | 47.8 | – | Paper | Code | 2020 | linked, not harvested | report | |
| 204 | Mask R-CNN (ResNeXt-101-FPN) | 36.7 | 59.5 | 38.9 | – | Paper | Code | 2017 | 42 of 140 ran · 98 unverified | report | ||||
| 205 | FoveaBox (ResNet-50-FPN, 600x600) | 36.0 | 55.2 | 37.9 | 18.6 | 39.4 | 50.5 | – | Paper | Code | 2019 | linked, not harvested | report | |
| 206 | FSAF (ResNet-50) | 35.9 | 55.0 | 37.9 | 19.8 | 39.6 | 48.2 | – | Paper | Code | 2019 | 0 of 8 ran · 8 unverified | report | |
| 207 | GHM-C + GHM-R (RetinaNet-FPN-ResNet-50, M=30) | 35.8 | 55.5 | 38.1 | 19.6 | 39.6 | 46.7 | – | Paper | Code | 2018 | 0 of 1 ran · 1 unverified | report | |
| 208 | Online Fg Bal. Sampling+Hard Negative Mining (ResNet-50) | 35.6 | 55.3 | – | Paper | Code | 2019 | linked, not harvested | report | |||||
| 209 | M2Det (ResNet-1o1, 320x320) | 34.1 | 53.7 | 15.9 | 39.5 | 49.3 | – | Paper | Code | 2018 | 1 of 1 ran · 0 unverified | report | ||
| 210 | Faster R-CNN (Res2Net-50) | 33.7 | 53.6 | 14 | 38.3 | 51.1 | – | Paper | Code | 2019 | 3 of 9 ran · 6 unverified | report | ||
| 211 | M2Det (VGG-16, 320x320) | 33.2 | 52.2 | 15 | 38.2 | 49.1 | – | Paper | Code | 2018 | 1 of 1 ran · 0 unverified | report | ||
| 212 | SOLQ (Swin-L, single scale) | 74.9 | 61.3 | 71.9 | – | Paper | Code | 2021 | linked, not harvested | report | ||||
| 213 | YOLOR-D6 (1280, single-scale, 31 fps) | 73.5 | 60.6 | 40.4 | 60.1 | 68.7 | – | Paper | Code | 2021 | linked, not harvested | report | ||
| 214 | EfficientDet-D7x (single-scale) | 73.4 | 59.0 | 40.0 | 58.0 | 67.9 | – | Paper | Code | 2019 | 11 of 70 ran · 59 unverified | report | ||
| 215 | YOLOR-P6 (1280, single-scale, 72 fps) | 70.6 | 57.4 | 37.4 | 57.3 | 65.2 | – | Paper | Code | 2021 | linked, not harvested | report | ||
| 216 | FocalNet-T (SRF, Cascade Mask R-CNN) | 70.1 | 55.8 | – | Paper | Code | 2022 | 0 of 2 ran · 2 unverified | report | |||||
| 217 | R3-CNN (ResNet-50-FPN, GRoIE) | 61.2 | 45.6 | 24.4 | – | Paper | Code | 2021 | linked, not harvested | report | ||||
| 218 | Mask R-CNN (HRNetV2p-W32, cascade) | 26.1 | 47.9 | – | Paper | Code | 2019 | 3 of 34 ran · 31 unverified | report | |||||
| 219 | Shift-T | 42.3 | – | Paper | Code | 2022 | 6 of 7 ran · 1 unverified | report | ||||||
| 220 | DyHead (ResNeXt-64x4d-101-DCN, multi scale) | 66.3 | – | Paper | Code | 2021 | 7 of 9 ran · 2 unverified | report |
All 220 rows shown. 220 link to a paper page on this site; 25 are marked as using additional training data in the archive. No GitHub stars are tracked; "Code" is the first repository the archive lists for the row. The archive carries no row tags, review links or community-submitted rows for this table; none are shown. archive 2025-07-28
Syntology Ran reads "N of M ran · U unverified": of the M code samples Syntology harvested from repositories linked to that row's paper (joined by arXiv id), N executed on a synthesized input and the other U = M−N are unverified (harvested, no recorded run). It counts code from repositories linked to that row's paper, not this result: the row's number was not reproduced and nothing here is a correctness claim. The other cell texts mean no graph line for the row: "linked, not harvested" (the archive links code, Syntology has not harvested it), "no code linked" (no code link in the archive), "not matched" (the row's paper URL matched no paper on this site). 147 rows have a graph line, from 66 distinct papers; 132 rows (58 papers) have at least one sample that ran. Counting each paper once: Syntology ran 785 of 1,678 samples; 893 unverified. Separately, 496 of those 1,678 are pointer-only (licence): the site points at that code rather than redistributing it, a licence property recorded for ran and unverified samples alike; each cell's tooltip carries the row's own pointer-only count. Read from the graph 2026-09-24. Per-sample status is on the paper page.
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