Papers › SetVAE: Learning Hierarchical Composition for Generative Modeling of Set-Structured Data

SetVAE: Learning Hierarchical Composition for Generative Modeling of Set-Structured Data

29 Mar 2021CVPR 2021 1arXiv:2103.15619archive 2025-07-28

Jinwoo Kim, Jaehoon Yoo, Juho Lee, Seunghoon Hong

Generative modeling of set-structured data, such as point clouds, requires reasoning over local and global structures at various scales. However, adopting multi-scale frameworks for ordinary sequential data to a set-structured data is nontrivial as it should be invariant to the permutation of its elements. In this paper, we propose SetVAE, a hierarchical variational autoencoder for sets. Motivated by recent progress in set encoding, we build SetVAE upon attentive modules that first partition the set and project the partition back to the original cardinality. Exploiting this module, our hierarchical VAE learns latent variables at multiple scales, capturing coarse-to-fine dependency of the set elements while achieving permutation invariance. We evaluate our model on point cloud generation task and achieve competitive performance to the prior arts with substantially smaller model capacity. We qualitatively demonstrate that our model generalizes to unseen set sizes and learns interesting subset relations without supervision. Our implementation is available at https://github.com/jw9730/setvae.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

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):

Code

Syntology 27 code samples harvested from 1 repository linked to this paper: 17 ran (of which 7 constructed an object rather than computing a result; 9 with no instrument failure: 2 honoured, 0 violated, 7 with no contract checked; 8 where Syntology's instrument failed) · 10 unverified. Where Syntology's instrument failed: 7 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 27 samples from 1 repository, 17 ran (9 with no instrument failure; 8 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.

Levels written on 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

jw9730/setvae officialmentioned in papermentioned on GitHubpytorchSyntology: 27 samples, 17 ran (8 instrument-failed) · 10 unverified report
jw9730/jw9730.github.io mentioned on GitHubSyntology: not harvested 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

27 samples harvested: 17 ran (of which 7 constructed an object rather than computing a result; 9 with no instrument failure: 2 honoured, 0 violated, 7 with no contract checked; 8 where Syntology's instrument failed) · 10 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.

2ran · honoured contract
7ran · our draft was wrong
1ran; the contract check could not run
7ran
10unverified

Licence: 0 of the 27 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 jw9730/setvae. “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 1 sample 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.

jw9730/setvae official repository 27 samples, 17 ran (8 instrument-failed) · 10 unverified
ChamferCriterion models/networks.py ran MIT (permissive) · 6787a72363072514 · report fixture: returned ChamferCriterion · level recorded 2026-09-22
CombinedCriterion models/networks.py ran MIT (permissive) · 2dbcaf1aa112361e · report fixture: returned CombinedCriterion · level recorded 2026-09-22
EMDCriterion models/networks.py ran MIT (permissive) · bed8a7b03b512335 · report fixture: returned EMDCriterion · level recorded 2026-09-22
InitialSet models/networks.py ran MIT (permissive) · ec91f5dff2cfdd95 · report fixture: returned InitialSet · level recorded 2026-09-22
MLP in 375 papers models/networks.py ran fingerprinted MIT (permissive) · 56ed680b39cc62e1 · report fixture: returned MLP · level recorded 2026-09-22
MaskedBatchNorm1d models/networks.py ran MIT (permissive) · 6ccac28dde2e0594 · report fixture: returned MaskedBatchNorm1d · level recorded 2026-09-22
ResidualAttention models/networks.py ran MIT (permissive) · d36dbf566a4f76e4 · report fixture: returned ResidualAttention · level recorded 2026-09-22
_chamfer_loss_v1 models/networks.py ran · our draft was wrong MIT (permissive) · c4401e3045759b3c · report fixture: signature output_set:Tensor/3/float, output_mask:Tensor/2/bool, target_set:Tensor/3/float, target_mask:Tensor/2/bool · returned Tensor · level recorded 2026-09-22
chamfer_loss in 6 papers models/networks.py ran; the contract check could not run (DEP_MISSING) MIT (permissive) · 42c489cac0ddf783 · report fixture: signature output_set:Tensor/3/float, output_mask:Tensor/2/bool, target_set:Tensor/3/float, target_mask:Tensor/2/bool, accelerate:bool/0/bool · returned Tensor · level recorded 2026-09-22
cube_to_sphere models/networks.py ran · our draft was wrong MIT (permissive) · 46b4d8dba0db14c3 · report fixture: signature cube:list/1/list · returned generator · level recorded 2026-09-22
get_mask in 119 papers
census: 9 implementations from 10 papers, 9 distinct outputs across 7 buckets, one shared input per bucket this sample: bucket 1 of 7, arg 1: rank 1, kind int, dtype int64
models/networks.py ran · our draft was wrong fingerprinted MIT (permissive) · 4e612cecc9647edc · report fixture: signature sizes:Tensor/1/int, max_size:int/0/int · returned Tensor · level recorded 2026-09-22
get_module in 37 papers models/networks.py ran · our draft was wrong MIT (permissive) · f5ca3e9834167675 · report fixture: signature gpu:int/0/int, module:callable/0/? · returned SimpleModel · level recorded 2026-09-22
get_pairwise_distance models/networks.py ran · our draft was wrong fingerprinted MIT (permissive) · 325df0adea9a09ec · report fixture: signature x:Tensor/2/float, p:int/0/int · returned Tensor · level recorded 2026-09-22
int_sin_m models/networks.py ran · honoured contract fingerprinted MIT (permissive) · 0c55dfd2474acb2e · report fixture: signature x:float/0/float, m:int/0/int · returned float · level recorded 2026-09-22
inverse_increasing models/networks.py ran · honoured contract MIT (permissive) · 6269a6d083ef6af9 · report fixture: signature func:callable/0/?, target:float/0/float, lower:float/0/float, upper:float/0/float, atol:float/0/float · returned float · level recorded 2026-09-22
masked_fill models/networks.py ran · our draft was wrong MIT (permissive) · 1d25c9b2a1df8aa1 · report fixture: signature tensor_bnc:Tensor/3/float, mask_bn:Tensor/2/bool, value:float/0/float · returned Tensor · level recorded 2026-09-22
sample_mask in 39 papers
census: 4 implementations from 5 papers, 4 distinct outputs on one shared input this sample: bucket 1 of 1, arg 1: rank 1, kind int, dtype int64
models/networks.py ran · our draft was wrong fingerprinted MIT (permissive) · f3ac4b1a18df2e8f · report fixture: signature sample_sizes:Tensor/1/int, max_size:int/0/int · returned Tensor · level recorded 2026-09-22
AttentiveBlock models/networks.py unverified MIT (permissive) · e76e125691f86272 · report
DecoderBlock in 30 papers models/networks.py unverified MIT (permissive) · 87658bdfa597bd52 · report
DeterministicNetwork models/networks.py unverified MIT (permissive) · 508dac2dcd3acee5 · report
ElementwiseMLP models/networks.py unverified MIT (permissive) · 3b42633e7587c43c · report
EncoderBlock in 23 papers models/networks.py unverified MIT (permissive) · 6d5bb50995b393f2 · report
ISAB models/networks.py unverified MIT (permissive) · 380e877aad19e438 · report
InducedNetwork models/networks.py unverified MIT (permissive) · 0c00bb082460dcd5 · report
SetVAE models/networks.py unverified MIT (permissive) · 195dea2fed61451a · report
emd_loss models/networks.py unverified MIT (permissive) · 06bb79fd238d6e69 · report
sphere_lattice models/networks.py unverified MIT (permissive) · d226bf1b07b03f2e · report

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 SetVAE (graph method: SetVAE). 27 bricks; individually on a fixture: 17 ran (of which 7 constructed an object rather than computing a result; 9 with no instrument failure: 2 honoured, 0 violated, 7 with no contract checked; 8 where Syntology's instrument failed) · 10 unverified. The assembly has not been run end-to-end.

  1. SetVAE entry · level 0 · unverified · 195dea2fed61451a
  2. ChamferCriterion loss · level 1 · ran · 6787a72363072514
  3. CombinedCriterion loss · level 1 · ran · 2dbcaf1aa112361e
  4. DecoderBlock model · level 0 · unverified · 87658bdfa597bd52
  5. DeterministicNetwork helper · level 0 · unverified · 508dac2dcd3acee5
  6. EMDCriterion loss · level 1 · ran · bed8a7b03b512335
  7. EncoderBlock model · level 0 · unverified · 6d5bb50995b393f2
  8. InitialSet model · level 1 · ran · ec91f5dff2cfdd95
  9. get_module helper · level 1 · ran · our draft was wrong · f5ca3e9834167675
  10. chamfer_loss loss · level 1 · ran; the contract check could not run (DEP_MISSING) · 42c489cac0ddf783
  11. emd_loss loss · level 0 · unverified · 06bb79fd238d6e69
  12. AttentiveBlock helper · level 0 · unverified · e76e125691f86272
  13. InducedNetwork helper · level 0 · unverified · 0c00bb082460dcd5
  14. ElementwiseMLP helper · level 0 · unverified · 3b42633e7587c43c
  15. ISAB helper · level 0 · unverified · 380e877aad19e438
  16. get_mask helper · level 1 · ran · our draft was wrong · 4e612cecc9647edc
  17. get_pairwise_distance helper · level 1 · ran · our draft was wrong · 325df0adea9a09ec
  18. sample_mask helper · level 1 · ran · our draft was wrong · f3ac4b1a18df2e8f
  19. sphere_lattice helper · level 0 · unverified · d226bf1b07b03f2e
  20. _chamfer_loss_v1 helper · level 1 · ran · our draft was wrong · c4401e3045759b3c
  21. ResidualAttention helper · level 1 · ran · d36dbf566a4f76e4
  22. MLP helper · level 1 · ran · 56ed680b39cc62e1
  23. masked_fill helper · level 1 · ran · our draft was wrong · 1d25c9b2a1df8aa1
  24. cube_to_sphere helper · level 1 · ran · our draft was wrong · 46b4d8dba0db14c3
  25. MaskedBatchNorm1d helper · level 1 · ran · 6ccac28dde2e0594
  26. int_sin_m helper · level 1 · ran · honoured contract · 0c55dfd2474acb2e
  27. inverse_increasing helper · level 1 · ran · honoured contract · 6269a6d083ef6af9

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="2103.15619") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Point Cloud Generation

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Point Cloud Generation ShapeNet Airplane SetVAE 1-NNA-CD 75.31 #4 of 5 Archive leaderboard report
Point Cloud Generation ShapeNet Airplane SetVAE MMD-CD 0.199 #4 of 5 Archive leaderboard report
Point Cloud Generation ShapeNet Car SetVAE 1-NNA-CD 59.66 #3 of 5 Archive leaderboard report
Point Cloud Generation ShapeNet Car SetVAE MMD-CD 0.88 #3 of 5 Archive leaderboard report
Point Cloud Generation ShapeNet Chair SetVAE 1-NNA-CD 58.76 #3 of 5 Archive leaderboard report
Point Cloud Generation ShapeNet Chair SetVAE MMD-CD 2.55 #3 of 5 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

AttentionHierarchical VAELinear LayerMulti-Head AttentionSet TransformerSoftmax

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