</> Code with Papers by Syntology

Browse › Computer Vision › Image Classification › ImageNet ReaL

Image Classification archive 2025-07-28

ImageNet ReaL Benchmark (Image Classification)

57 rows 56 with code listed 4 metrics

Image Classification is a fundamental task in vision recognition that aims to understand and categorize an image as a whole under a specific label. Unlike object detection, which involves classification and location of multiple objects within an image, image classification typically pertains to single-object images. When the classification becomes highly detailed or reaches instance-level, it is often referred to as image retrieval, which also involves finding similar images in a large database.

Source: Metamorphic Testing for Object Detection Systems

The archive carries no text for this table; the description above is the archive's text for the task Image Classification. archive 2025-07-28

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: Accuracy (higher is better), Params (lower is better), Top 1 Accuracy (higher is better), Number of params (lower is better). Points are placed at the row's paper date; 57 of 57 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 Baseline (ViT-G/14) 91.78% ✓ Paper Code 2022 5 of 17 ran · 12 unverified report
2 ViTAE-H (MAE, 512) 91.2%644M – Paper Code 2022 15 of 25 ran · 10 unverified report
3 Model soups (ViT-G/14) 91.20%1843M ✓ Paper Code 2022 5 of 17 ran · 12 unverified report
4 Meta Pseudo Labels (EfficientNet-B6-Wide) 91.12% ✓ Paper Code 2020 5 of 14 ran · 9 unverified report
5 MAWS (ViT-6.5B) 91.1% ✓ Paper Code 2023 linked, not harvested report
6 TokenLearner L/8 (24+11) 91.05%460M ✓ Paper Code 2021 3 of 3 ran · 0 unverified report
7 Model soups (BASIC-L) 91.03%2440M ✓ Paper Code 2022 5 of 17 ran · 12 unverified report
8 Meta Pseudo Labels (EfficientNet-L2) 91.02% ✓ Paper Code 2020 5 of 14 ran · 9 unverified report
9 MAWS (ViT-2B) 90.9% ✓ Paper Code 2023 linked, not harvested report
10 FixEfficientNet-L2 90.9%480M ✓ Paper Code 2020 linked, not harvested report
11 ViT-G/14 90.81% ✓ Paper Code 2021 linked, not harvested report
12 MAWS (ViT-H) 90.8% ✓ Paper Code 2023 linked, not harvested report
13 SWAG (RegNetY 128GF) 90.7% ✓ Paper Code 2022 linked, not harvested report
14 CvT-W24 (384 res, ImageNet-22k pretrain) 90.6%87.7%277M ✓ Paper Code 2021 29 of 47 ran · 18 unverified report
15 VOLO-D5 90.6% – Paper Code 2021 1 of 6 ran · 5 unverified report
16 EfficientNet-L2 90.55%480M ✓ Paper Code 2019 5 of 24 ran · 19 unverified report
17 BiT-L 90.54%928M ✓ Paper Code 2019 3 of 10 ran · 7 unverified report
18 VOLO-D4 90.5% – Paper Code 2021 1 of 6 ran · 5 unverified report
19 CAIT-M36-448 90.2% – Paper Code 2021 5 of 11 ran · 6 unverified report
20 Mixer-H/14- 448 (JFT-300M pre-train) 90.18%409M ✓ Paper Code 2021 106 of 134 ran · 28 unverified report
21 FixEfficientNet-B8 90.0%87M – Paper Code 2020 linked, not harvested report
22 SEER (RegNet10B) 89.8%10000M ✓ Paper Code 2022 linked, not harvested report
23 FixResNeXt-101 32x48d 89.73%829M ✓ Paper Code 2019 0 of 2 ran · 2 unverified report
24 DeiT-B-384 89.3%86M – Paper Code 2020 12 of 19 ran · 7 unverified report
25 BiT-M 89.02% ✓ Paper Code 2019 3 of 10 ran · 7 unverified report
26 DeiT-B 88.7%86M – Paper Code 2020 12 of 19 ran · 7 unverified report
27 Assemble-ResNet152 88.65% – Paper Code 2020 1 of 11 ran · 10 unverified report
28 CeiT-S (384 finetune res) 88.1% – Paper Code 2021 9 of 11 ran · 2 unverified report
29 Sequencer2D-L 87.9 – Paper Code 2022 4 of 9 ran · 5 unverified report
30 Mixer-H/14 (JFT-300M pre-train) 87.86%409M ✓ Paper Code 2021 106 of 134 ran · 28 unverified report
31 Assemble ResNet-50 87.82% – Paper Code 2020 1 of 11 ran · 10 unverified report
32 NASNet-A Large 87.56% – Paper Code 2017 0 of 1 ran · 1 unverified report
33 LeViT-384 87.5% – Paper Code 2021 23 of 30 ran · 7 unverified report
34 CeiT-S 87.3% – Paper Code 2021 9 of 11 ran · 2 unverified report
35 LeViT-256 86.9% – Paper Code 2021 23 of 30 ran · 7 unverified report
36 DeiT-S 86.8%22M – Paper Code 2020 12 of 19 ran · 7 unverified report
37 ResNet-152x2-SAM 86.4% – Paper Code 2021 linked, not harvested report
38 LeViT-192 85.8% – Paper Code 2021 23 of 30 ran · 7 unverified report
39 ResNet50 (A1) 85.7%25M – Paper Code 2021 0 of 3 ran · 3 unverified report
40 LeViT-128 85.6% – Paper Code 2021 23 of 30 ran · 7 unverified report
41 ResMLP-36 85.6%45M – Paper Code 2021 2 of 7 ran · 5 unverified report
42 ResMLP-24 85.3%30M – Paper Code 2021 2 of 7 ran · 5 unverified report
43 ViT-B/16-SAM 85.2% – Paper Code 2021 linked, not harvested report
44 ResMLP-12 84.6%15M – Paper Code 2021 2 of 7 ran · 5 unverified report
45 Mixer-B/8-SAM 84.4% – Paper Code 2021 linked, not harvested report
46 kNN-CLIP 84% ✓ Paper – 2022 no code linked report
47 CeiT-T 83.6% – Paper Code 2021 9 of 11 ran · 2 unverified report
48 LeViT-128S 82.6% – Paper Code 2021 23 of 30 ran · 7 unverified report
49 DeiT-Ti 82.1%5M – Paper Code 2020 12 of 19 ran · 7 unverified report
50 NASNet-A Mobile 81.15% – Paper Code 2017 0 of 1 ran · 1 unverified report
51 VGG-16 BN 80.60% – Paper Code 2014 12 of 122 ran · 110 unverified report
52 VGG-16 79.01% – Paper Code 2014 12 of 122 ran · 110 unverified report
53 AlexNet 62.88% – Paper Code 2012 linked, not harvested report
54 ViT-L @384 (DeiT III, 21k) 87.7%304M – Paper Code 2022 linked, not harvested report
55 ViT-H @224 (DeiT III, 21k) 87.2%632M – Paper Code 2022 linked, not harvested report
56 ViT-L @224 (DeiT III, 21k) 87.0% – Paper Code 2022 linked, not harvested report
57 ResMLP-B24/8 (22k) 84.4% – Paper Code 2021 2 of 7 ran · 5 unverified report

All 57 rows shown. 57 link to a paper page on this site; 21 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). 41 rows have a graph line, from 20 distinct papers; 37 rows (17 papers) have at least one sample that ran. Counting each paper once: Syntology ran 240 of 506 samples; 266 unverified. Separately, 54 of those 506 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