Browse › Computer Vision › Image Generation › ImageNet 64x64

Image Generation archive 2025-07-28

ImageNet 64x64 Benchmark (Image Generation)

65 rows 52 with code listed 5 metrics

Image Generation (synthesis) is the task of generating new images from an existing dataset.

In this section, you can find state-of-the-art leaderboards for unconditional generation. For conditional generation, and other types of image generations, refer to the subtasks.

The archive carries no text for this table; the description above is the archive's text for the task Image Generation. 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: FID (lower is better), Inception Score (higher is better), KID (lower is better). Not inferred (points only, no best-so-far line): Bits per dim, NFE. Points are placed at the row's paper date; 65 of 65 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 SIMS 0.92126 – Paper – 2024 no code linked report
2 EDM2-S+DDO 0.9763 – Paper Code 2025 linked, not harvested report
3 Uni-Instruct 1.021 – Paper – 2025 no code linked report
4 SiDA-EDM 1.111 – Paper Code 2024 2 of 3 ran · 1 unverified report
5 GDD-I 1.161 – Paper Code 2024 linked, not harvested report
6 PaGoDA 1.21176.47 – Paper Code 2024 15 of 17 ran · 2 unverified report
7 DisCo-Diff 1.22 – Paper Code 2024 6 of 6 ran · 0 unverified report
8 RIN 1.23 – Paper Code 2022 21 of 29 ran · 8 unverified report
9 GDD 1.421 – Paper Code 2024 linked, not harvested report
10 SCT 1.472 – Paper Code 2024 linked, not harvested report
11 CDM 1.48 – Paper – 2021 no code linked report
12 StyleGAN-XL 1.511 – Paper Code 2022 14 of 19 ran · 5 unverified report
13 SiD 1.5241 – Paper Code 2024 9 of 15 ran · 6 unverified report
14 TCM 1.622 – Paper – 2024 no code linked report
15 ECM-XL 1.672 – Paper Code 2024 2 of 3 ran · 1 unverified report
16 CAF 1.69262.03 – Paper Code 2024 3 of 3 ran · 0 unverified report
17 CTM 1.73264.29 – Paper Code 2023 14 of 18 ran · 4 unverified report
18 ADM (dropout) 2.07 – Paper Code 2021 27 of 50 ran · 23 unverified report
19 LEGO 2.1678.7 – Paper Code 2023 11 of 17 ran · 6 unverified report
20 TarFlow 2.92.99 – Paper Code 2024 1 of 4 ran · 3 unverified report
21 Improved DDPM 2.923.53 – Paper Code 2021 9 of 11 ran · 2 unverified report
22 2-rectified flow++ (NFE=2) 3.64 – Paper Code 2024 1 of 1 ran · 0 unverified report
23 2-rectified flow++ (NFE=1) 4.31 – Paper Code 2024 1 of 1 ran · 0 unverified report
24 CD (Diffusion + Distillation, NFE=2) 4.702 – Paper Code 2023 28 of 57 ran · 29 unverified report
25 CD (Diffusion + Distillation, NFE=1) 6.201 – Paper Code 2023 28 of 57 ran · 29 unverified report
26 CT (Direct Generation, NFE=2) 11.12 – Paper Code 2023 28 of 57 ran · 29 unverified report
27 CT (Direct Generation, NFE=1) 13.01 – Paper Code 2023 28 of 57 ran · 29 unverified report
28 FM 14.453.31 – Paper Code 2022 1 of 4 ran · 3 unverified report
29 CLR-GAN 20.27 – Paper Code 2024 linked, not harvested report
30 PGMGAN 21.73 – Paper Code 2021 2 of 3 ran · 1 unverified report
31 GLIDE + CLIP + CLS + CLS-FREE 29.18434.9523.766 – Paper – 2022 no code linked report
32 GLIDE + CLS-FREE 29.21925.9265.325 – Paper – 2022 no code linked report
33 GLIDE + CLIP 30.46225.0176.174 – Paper – 2022 no code linked report
34 GLIDE + CLS 30.87122.077 – Paper – 2022 no code linked report
35 NFDM 3.2 – Paper Code 2024 linked, not harvested report
36 BSI 3.22 – Paper Code 2025 linked, not harvested report
37 Efficient-VDVAE 3.30 (different downsampling) – Paper Code 2022 linked, not harvested report
38 DenseFlow-74-10 3.35 (different downsampling) – Paper Code 2021 linked, not harvested report
39 NDM 3.35 – Paper – 2023 no code linked report
40 VDM 3.40 – Paper Code 2021 29 of 38 ran · 9 unverified report
41 Combiner-Axial 3.42 – Paper Code 2021 1 of 3 ran · 2 unverified report
42 Routing Transformer 3.43 – Paper Code 2020 3 of 3 ran · 0 unverified report
43 Sparse Transformer 59M (strided) 3.44 – Paper Code 2019 5 of 6 ran · 1 unverified report
44 MRCNF 3.44 – Paper Code 2021 linked, not harvested report
45 Hourglass 3.44 – Paper Code 2021 2 of 4 ran · 2 unverified report
46 Combiner-Mixture 3.504 – Paper Code 2021 1 of 3 ran · 2 unverified report
47 SPN 3.52 – Paper – 2018 no code linked report
48 Very Deep VAE 3.52 – Paper Code 2020 8 of 9 ran · 1 unverified report
49 PixelCNN 3.57 – Paper – 2016 no code linked report
50 Gated PixelCNN (van den Oord et al., [2016c]) 3.57 – Paper Code 2016 4 of 14 ran · 10 unverified report
51 Performer (12 layers) 3.636 – Paper Code 2020 9 of 16 ran · 7 unverified report
52 Flow++ 3.69 – Paper Code 2019 0 of 9 ran · 9 unverified report
53 MaCow (Var) 3.69 – Paper Code 2019 0 of 3 ran · 3 unverified report
54 Parallel Multiscale 3.7 – Paper – 2017 no code linked report
55 MALI 3.71 – Paper Code 2021 2 of 4 ran · 2 unverified report
56 Reformer (12 layers) 3.710 – Paper Code 2020 6 of 8 ran · 2 unverified report
57 Performer (6 layers) 3.719 – Paper Code 2020 9 of 16 ran · 7 unverified report
58 Reformer (6 layers) 3.740 – Paper Code 2020 6 of 8 ran · 2 unverified report
59 MaCow (Unf) 3.75 – Paper Code 2019 0 of 3 ran · 3 unverified report
60 Residual Flow 3.757 – Paper Code 2019 5 of 8 ran · 3 unverified report
61 Glow (Kingma and Dhariwal, 2018) 3.81 – Paper Code 2018 75 of 129 ran · 54 unverified report
62 Axial Transformer (6 layers) 4.032 – Paper Code 2019 3 of 3 ran · 0 unverified report
63 Logsparse (6 layers) 4.351 – Paper Code 2019 linked, not harvested report
64 CTM (NFE 1) 170.38 – Paper Code 2023 14 of 18 ran · 4 unverified report
65 GLIDE +CLS 7.952 – Paper – 2022 no code linked report

All 65 rows shown. 65 link to a paper page on this site; 0 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 32 distinct papers; 38 rows (30 papers) have at least one sample that ran. Counting each paper once: Syntology ran 318 of 517 samples; 199 unverified. Separately, 102 of those 517 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