Papers › NVAE: A Deep Hierarchical Variational Autoencoder
NVAE: A Deep Hierarchical Variational Autoencoder
Arash Vahdat, Jan Kautz
Normalizing flows, autoregressive models, variational autoencoders (VAEs), and deep energy-based models are among competing likelihood-based frameworks for deep generative learning. Among them, VAEs have the advantage of fast and tractable sampling and easy-to-access encoding networks. However, they are currently outperformed by other models such as normalizing flows and autoregressive models. While the majority of the research in VAEs is focused on the statistical challenges, we explore the orthogonal direction of carefully designing neural architectures for hierarchical VAEs. We propose Nouveau VAE (NVAE), a deep hierarchical VAE built for image generation using depth-wise separable convolutions and batch normalization. NVAE is equipped with a residual parameterization of Normal distributions and its training is stabilized by spectral regularization. We show that NVAE achieves state-of-the-art results among non-autoregressive likelihood-based models on the MNIST, CIFAR-10, CelebA 64, and CelebA HQ datasets and it provides a strong baseline on FFHQ. For example, on CIFAR-10, NVAE pushes the state-of-the-art from 2.98 to 2.91 bits per dimension, and it produces high-quality images on CelebA HQ. To the best of our knowledge, NVAE is the first successful VAE applied to natural images as large as 256×256 pixels. The source code is available at https://github.com/NVlabs/NVAE .
In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.
Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
For agents, on Syntology's MCP service (how to connect):
get_harvested_code_for_paper(arxiv_id="2007.03898")get_code(code_sha256="…")with a sample'scode_sha256from the list belowget_citation_path(paper_1="2007.03898", paper_2="…")with another paper's arXiv id or titleget_concepts_for_paper(arxiv_id="2007.03898")
Code
Syntology 41 code samples harvested from 3 repositories linked to this paper: 26 ran (of which 15 constructed an object rather than computing a result; 21 with no instrument failure: 3 honoured, 0 violated, 18 with no contract checked; 5 where Syntology's instrument failed) · 15 unverified. Where Syntology's instrument failed: 3 where our contract draft was wrong; 2 where the contract check could not run; failures of Syntology's instrument, not of the code.
By repository: official repository: 23 samples from 1 repository, 9 ran (6 with no instrument failure; 3 where Syntology's instrument failed); community (archive-listed): 18 samples from 2 repositories, 17 ran (15 with no instrument failure; 2 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.
No sample is linked to this paper from 3 of its repositories that Syntology harvested for other papers:
- NVlabs/LSGM, community (archive-listed): harvested for another paper, listed under the repository
- NVlabs/VAEBM, community (archive-listed): harvested for another paper, listed under the repository
- lagergren-lab/miselbo, community (archive-listed): harvested for another paper, listed under the repository
Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.
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.
A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.
Code Syntology ran Syntology
41 samples harvested: 26 ran (of which 15 constructed an object rather than computing a result; 21 with no instrument failure: 3 honoured, 0 violated, 18 with no contract checked; 5 where Syntology's instrument failed) · 15 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.
Licence: 23 of the 41 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; the list below is grouped by repository, each group headed by its name and what it is to the paper. “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. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. For 10 samples that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.
Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (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 are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. 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 six 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; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. 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.
A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.
3b69906c3f162603 · report
fixture: returned DiscMixLogistic · level recorded 2026-09-22 47f8add81ef809e7 · report
fixture: returned SE · level recorded 2026-09-22 b014ff059854d82e · report
fixture: returned SyncBatchNormSwish · level recorded 2026-09-22 4ca5ffe947c454b7 · report
fixture: signature dataset:str/0/str · returned int · level recorded 2026-09-22 417c413f2e89d954 · report
fixture: signature cell_type:str/0/str · returned int · level recorded 2026-09-22 ec8681cb9635d8c2 · report
fixture: signature num_scales:int/0/int, num_groups_per_scale:int/0/int, is_adaptive:bool/0/bool, divider:int/0/int, minimum_groups:int/0/int · returned list · level recorded 2026-09-22 9d358b42347bc9fc · report
fixture: signature z1:Tensor/2/float, logit_pi:Tensor/3/float, mu:Tensor/3/float, log_s:Tensor/3/float, log_a:Tensor/3/float, b:Tensor/3/float · returned tuple · level recorded 2026-09-22 509ea43acade44e9 · report
fixture: returned Tensor · level recorded 2026-09-22 census: 18 implementations from 25 papers, 10 distinct outputs across 6 buckets, one shared input per bucket this sample: bucket 2 of 6, arg 1: rank 2, kind int, dtype int64 model.py ran · our draft was wrong licence not identified · pointer only ·
b95bc8179cca09d6 · report
fixture: returned Tensor · level recorded 2026-09-22 213dfd5e4d31949b · report
ed8aab6b16f71a33 · report
095d98d4a1f55bfd · report
a674092cc931b81d · report
05cbbc54fd34ad83 · report
464aa3f0a27bcf6b · report
c4d8db6435b9c07a · report
277ed0e40f07a938 · report
dd83a8a0b04c47fd · report
b8e68f7d505c301c · report
e79daa47e7ea4581 · report
fixture: returned DecCombinerCell · level recorded 2026-09-22 dc14f24bd29f3b0d · report
fixture: returned DiscMixLogistic · level recorded 2026-09-22 fd98ee5b850457e9 · report
fixture: returned EncCombinerCell · level recorded 2026-09-22 ea5de3e39f202a06 · report
fixture: returned MaskedConv2d · level recorded 2026-09-22 2bf1bbe0e0e1f51c · report
fixture: returned NFBlock · level recorded 2026-09-22 a82d266ee79c8089 · report
fixture: returned NFCell · level recorded 2026-09-22 c9ab03bfda9403fb · report
fixture: returned Normal · level recorded 2026-09-22 21cbfdf47981afd2 · report
fixture: returned ResidualCellDecoder · level recorded 2026-09-22 1d1460d11268fcbe · report
fixture: returned ResidualCellEncoder · level recorded 2026-09-22 852119a74dcd01db · report
fixture: returned SE · level recorded 2026-09-22 90b219a8fe64989e · report
fixture: returned SkipDown · level recorded 2026-09-22 a8aa4571eb4d2e83 · report
fixture: returned SkipUp · level recorded 2026-09-22 cb77f2d04693965c · report
fixture: signature logits:Tensor/2/float, temperature:float/0/float · returned Tensor · level recorded 2026-09-22 081e9e26fd1942f7 · report
fixture: signature x:Tensor/2/float, n:float/0/float · returned Tensor · level recorded 2026-09-22 5f0c5cf7f773ac3b · report
fixture: signature x:EagerTensor/4/float · returned EagerTensor · level recorded 2026-09-24 c5a084857e4bb9e0 · report
fixture: signature layer_dict:dict/0/?, name:str/0/str · returned SymbolicTensor · level recorded 2026-09-24 census: 14 implementations from 37 papers, 2 distinct outputs on one shared input modules.py ran MIT (permissive) ·
248788c293d8eccd · report
fixture: signature layer_dict:dict/0/?, name:str/0/str · returned EagerTensor · level recorded 2026-09-24 Method as code Syntology
What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.
1 assembly on this paper: 1 harvested, 0 generated.
harvested Assembly for NVAE (graph method: NVAE). 15 bricks; individually on a fixture: 14 ran (of which 12 constructed an object rather than computing a result; 12 with no instrument failure: 0 honoured, 0 violated, 12 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified. The assembly has not been run end-to-end.
- AutoEncoder entry · level 0 · unverified ·
2b09442d4473a79b - DecCombinerCell model · level 1 · ran ·
e79daa47e7ea4581 - DiscMixLogistic model · level 1 · ran ·
dc14f24bd29f3b0d - EncCombinerCell model · level 1 · ran ·
fd98ee5b850457e9 - MaskedConv2d helper · level 1 · ran ·
ea5de3e39f202a06 - NFBlock model · level 1 · ran ·
2bf1bbe0e0e1f51c - Normal model · level 1 · ran ·
c9ab03bfda9403fb - ResidualCellDecoder model · level 1 · ran ·
21cbfdf47981afd2 - ResidualCellEncoder model · level 1 · ran ·
1d1460d11268fcbe - gumbel_sampling helper · level 1 · ran · our draft was wrong ·
cb77f2d04693965c - NFCell model · level 1 · ran ·
a82d266ee79c8089 - soft_clamp helper · level 1 · ran · our draft was wrong ·
081e9e26fd1942f7 - SE helper · level 1 · ran ·
852119a74dcd01db - SkipUp helper · level 1 · ran ·
a8aa4571eb4d2e83 - SkipDown helper · level 1 · ran ·
90b219a8fe64989e
Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2007.03898") returns these assemblies with each brick's role and wiring (how to connect).
Tasks
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Image Generation | CIFAR-10 | NVAE w/ flow | FID | 32.53 | #70 of 78 | Archive leaderboard | report |
| Image Generation | CelebA 256x256 | NVAE w/ flow | bpd | 0.70 | #5 of 17 | Archive leaderboard | report |
| Image Generation | FFHQ 256 x 256 | NVAE w/ flow | bits/dimension | 0.69 | #51 of 51 | Archive leaderboard | report |
| Image Generation | ImageNet 32x32 | NVAE w/ flow | bpd | 3.92 | #29 of 35 | 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: NVAE Encoder Residual Cell, NVAE Generative Residual Cell, Residual Normal Distribution
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