Papers › Scalable Adaptive Computation for Iterative Generation

Scalable Adaptive Computation for Iterative Generation

22 Dec 2022arXiv:2212.11972archive 2025-07-28

Allan Jabri, David Fleet, Ting Chen

Natural data is redundant yet predominant architectures tile computation uniformly across their input and output space. We propose the Recurrent Interface Networks (RINs), an attention-based architecture that decouples its core computation from the dimensionality of the data, enabling adaptive computation for more scalable generation of high-dimensional data. RINs focus the bulk of computation (i.e. global self-attention) on a set of latent tokens, using cross-attention to read and write (i.e. route) information between latent and data tokens. Stacking RIN blocks allows bottom-up (data to latent) and top-down (latent to data) feedback, leading to deeper and more expressive routing. While this routing introduces challenges, this is less problematic in recurrent computation settings where the task (and routing problem) changes gradually, such as iterative generation with diffusion models. We show how to leverage recurrence by conditioning the latent tokens at each forward pass of the reverse diffusion process with those from prior computation, i.e. latent self-conditioning. RINs yield state-of-the-art pixel diffusion models for image and video generation, scaling to 1024X1024 images without cascades or guidance, while being domain-agnostic and up to 10X more efficient than 2D and 3D U-Nets.

PaperPDFCodeCode 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.

Syntology holds a reference implementation of Recurrent Interface Networks (RINs), generated from the paper text (not the authors' code), at level 3 on Syntology's ladder: V2 “two implementations GENERATED from the paper's signature and spec agree numerically (1e-6) on fixture inputs” “neither of the two ever read the authors' own code, so V2 measures how unambiguously the paper specifies the method, not fidelity to the authors' implementation (measured 2026-09-11, commit 75eaa77: cross_checked_by is a second LLM on 2,628 of 2,831 served samples)”; V3 “paper-derived property tests pass on both of those same two generations.” (quoted from Syntology's server card). No rung compares this code against the authors' implementation. For agents: get_reference_implementation(name='Recurrent Interface Networks (RINs)') (how to connect); the free check have('2212.11972') needs no token.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 29 code samples harvested from 2 repositories linked to this paper: 21 ran (of which 9 constructed an object rather than computing a result; 18 with no instrument failure: 8 honoured, 0 violated, 10 with no contract checked; 3 where Syntology's instrument failed) · 8 unverified. Where Syntology's instrument failed: 2 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: 18 samples from 1 repository, 13 ran (10 with no instrument failure; 3 where Syntology's instrument failed); community (archive-listed): 11 samples from 1 repository, 8 ran (8 with no instrument failure; 0 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.

google-research/pix2seq officialmentioned in papermentioned on GitHubtfSyntology: 18 samples, 13 ran (3 instrument-failed) · 5 unverified report
lucidrains/recurrent-interface-network-pytorch mentioned on GitHubpytorchSyntology: 11 samples, 8 ran (0 instrument-failed) · 3 unverified 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

29 samples harvested: 21 ran (of which 9 constructed an object rather than computing a result; 18 with no instrument failure: 8 honoured, 0 violated, 10 with no contract checked; 3 where Syntology's instrument failed) · 8 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.

8ran · honoured contract
2ran · our draft was wrong
1ran; the contract check could not run
10ran
8unverified

Licence: 1 of the 29 samples is 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 2 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. 1 sample whose link to this paper records no repository is listed under one of this paper's own repositories, because identical code was harvested from it for another 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 3 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.

google-research/pix2seq official repository 18 samples, 13 ran (3 instrument-failed) · 5 unverified
DropPath in 75 papers
census: 17 implementations from 17 papers, 8 distinct outputs across 3 buckets, one shared input per bucket
architectures/transformers.py ran Apache-2.0 (permissive) · 80161ca59f374d2a · report fixture: returned DropPath · level recorded 2026-09-22
FeedForwardLayer in 9 papers architectures/transformers.py ran Apache-2.0 (permissive) · b9ad5f9347bcd710 · report fixture: returned FeedForwardLayer · level recorded 2026-09-22
add_vis_pos_emb architectures/transformers.py ran · honoured contract Apache-2.0 (permissive) · fbc958c1295b5f01 · report fixture: signature self:MockModel/0/?, pos_encoding:str/0/str, n_rows:int/0/int, n_cols:int/0/int, dim:int/0/int, name_prefix:str/0/str, initializer:NoneType/0/?, return_only:bool/0/bool, normalization_max:float/0/float · returned SymbolicTensor · level recorded 2026-09-22
get_1d_position_codes architectures/transformers.py ran · honoured contract fingerprinted Apache-2.0 (permissive) · cbbd048907025ae1 · report fixture: signature seqlen:int/0/int, out_dim:int/0/int, normalization_max:float/0/float · returned EagerTensor · level recorded 2026-09-22
get_2d_position_codes architectures/transformers.py ran · honoured contract fingerprinted Apache-2.0 (permissive) · 554d30594bb18ba4 · report fixture: signature height:int/0/int, width:int/0/int, out_dim:int/0/int, normalization_max:float/0/float · returned EagerTensor · level recorded 2026-09-22
get_angles in 21 papers
census: 3 implementations from 7 papers, 3 distinct outputs across 3 buckets, one shared input per bucket
architectures/transformers.py ran · honoured contract Apache-2.0 (permissive) · c1a5330036a08dfa · report fixture: signature pos:int/0/int, i:int/0/int, dim:int/0/int · returned EagerTensor · level recorded 2026-09-22
get_ar_mask architectures/transformers.py ran · honoured contract Apache-2.0 (permissive) · 508ee551fb33fedc · report fixture: signature seq_len:int/0/int, dtype:DType/0/? · returned EagerTensor · level recorded 2026-09-22
get_chunk_ar_mask architectures/transformers.py ran · honoured contract Apache-2.0 (permissive) · 3e4ac71a6f12be9d · report fixture: signature seq_len:int/0/int, chunk_size:int/0/int, dtype:DType/0/? · returned EagerTensor · level recorded 2026-09-22
get_layer_config architectures/transformers.py ran · honoured contract Apache-2.0 (permissive) · 90a86ebf31046958 · report fixture: signature layers_str:str/0/str · returned list · level recorded 2026-09-22
get_variable_initializer architectures/transformers.py ran · our draft was wrong Apache-2.0 (permissive) · 37888d00e9d8c4ab · report fixture: signature name:NoneType/0/? · returned TruncatedNormal · level recorded 2026-09-22
kronecker_product in 9 papers
census: 2 implementations from 6 papers, 1 distinct output on one shared input
architectures/transformers.py ran · honoured contract fingerprinted Apache-2.0 (permissive) · e37890b1d73b280c · report fixture: signature mat1:EagerTensor/2/float, mat2:EagerTensor/2/float · returned EagerTensor · level recorded 2026-09-22
positional_encoding in 63 papers
census: 5 implementations from 6 papers, 4 distinct outputs on one shared input
architectures/transformers.py ran; the contract check could not run (RAISES) fingerprinted Apache-2.0 (permissive) · 7a5b750c9c8d5e81 · report fixture: signature coords:EagerTensor/2/int, dim:int/0/int · returned SymbolicTensor · level recorded 2026-09-22
suffix_id architectures/transformers.py identical code Syntology harvested from this repository for another paper; this paper's link records no repository, path or licence of its own ran · our draft was wrong fingerprinted licence of this copy not recorded · d1c561539ef11926 · report fixture: signature i:int/0/int · returned str · level recorded 2026-09-22
FIT architectures/transformers.py unverified Apache-2.0 (permissive) · e147f983e472692f · report
MLP in 375 papers architectures/transformers.py unverified Apache-2.0 (permissive) · e2256176baee1426 · report
TransformerDecoderLayer in 45 papers architectures/transformers.py unverified Apache-2.0 (permissive) · f14aac70f9859dd8 · report
TransformerEncoder in 96 papers architectures/transformers.py unverified Apache-2.0 (permissive) · b019a0895599116d · report
TransformerEncoderLayer in 90 papers architectures/transformers.py unverified Apache-2.0 (permissive) · e730a175de3303f8 · report
lucidrains/recurrent-interface-network-pytorch community (archive-listed) 11 samples, 8 ran (0 instrument-failed) · 3 unverified
Attend in 6 papers rin_pytorch/rin_pytorch.py ran MIT (permissive) · df4d0ae2244e374f · report fixture: returned Attend · level recorded 2026-09-22
Attention in 344 papers
census: 13 implementations from 12 papers, 10 distinct outputs across 4 buckets, one shared input per bucket
rin_pytorch/rin_pytorch.py ran MIT (permissive) · 305ed1d536ee99b2 · report fixture: returned Attention · level recorded 2026-09-22
FeedForward in 198 papers rin_pytorch/rin_pytorch.py ran MIT (permissive) · 1e9eca95c178f7c9 · report fixture: returned FeedForward · level recorded 2026-09-22
LayerNorm in 121 papers
census: 5 implementations from 5 papers, 5 distinct outputs across 3 buckets, one shared input per bucket
rin_pytorch/rin_pytorch.py ran MIT (permissive) · d070db73849f9743 · report fixture: returned LayerNorm · level recorded 2026-09-22
LearnedSinusoidalPosEmb rin_pytorch/rin_pytorch.py ran · metamorphic tier: invariant fingerprinted MIT (permissive) · e4a1970252d572fd · report fixture: signature x:Tensor/1/float · returned Tensor · level recorded 2026-09-22
LinearAttention in 20 papers rin_pytorch/rin_pytorch.py ran MIT (permissive) · f8753f538bdf8090 · report fixture: returned LinearAttention · level recorded 2026-09-22
MultiHeadedRMSNorm rin_pytorch/rin_pytorch.py ran MIT (permissive) · 8535fbe5734cf1d3 · report fixture: returned MultiHeadedRMSNorm · level recorded 2026-09-22
PEG in 5 papers rin_pytorch/rin_pytorch.py ran MIT (permissive) · 2090cfddb3b4ea08 · report fixture: returned PEG · level recorded 2026-09-22
RIN rin_pytorch/rin_pytorch.py unverified MIT (permissive) · 5fdaf6caf1863ed1 · report
RINBlock rin_pytorch/rin_pytorch.py unverified MIT (permissive) · 89c11df753e06cb9 · report
Sequential rin_pytorch/rin_pytorch.py unverified MIT (permissive) · d6fde31f78c4502f · 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. For a generated assembly, an independent peer is a second implementation generated separately from the same paper text; neither implementation read the authors' code. At level 2 the two agreed numerically on the fixture's inputs; level 3 is recorded when property tests derived from the paper also passed on both, and each level 3 card prints how many its report lists. 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. A generated assembly's label was set when its code was generated for the method, and no model judged it; its card prints it as "by construction" or "not judged by a model". Siblings are assemblies lifted from this paper in the same batch for the same method: alternative readings kept side by side; none replaces or supersedes another. Where an assembly's phrase has no constructed figure, a brick that ran is not among the harvested samples this build reads, and whether a brick's run only constructed an object is read from that brick's harvested sample, so the figure was not counted; it does not say there were none.

3 assemblies on this paper: 2 harvested, 1 generated, listed in 2 groups (siblings lifted in one batch form one group).

generated Assembly for Recurrent Interface Networks (RINs) (graph method: Recurrent Interface Networks (RINs)), generated from the paper text, not the repository's code. 1 brick; individually on a fixture: 1 ran (1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified. Ran end-to-end against an independent peer (level 3; 4 property tests passed on both).

label matches by construction: the code was generated for this method (not judged by a model)

  1. multihead_attention entry · level 3 · ran · honoured contract · a7992dbe74aef33a483b9d1ea2882e116729c49b72d3ccbed3fc8cb46db87870

2 sibling assemblies for Recurrent Interface Networks (RINs) (harvested), lifted in the same batch on 2026-09-09: alternative readings kept side by side; none replaces or supersedes another.

harvested Assembly for Recurrent Interface Networks (RINs) (graph method: Recurrent Interface Networks (RINs)). 18 bricks; individually on a fixture: 13 ran (of which 2 constructed an object rather than computing a result; 10 with no instrument failure: 8 honoured, 0 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 5 unverified. The assembly has not been run end-to-end.

one of 2 siblings

  1. FIT entry · level 0 · unverified · e147f983e472692f
  2. TransformerDecoderLayer model · level 0 · unverified · f14aac70f9859dd8
  3. TransformerEncoder helper · level 0 · unverified · b019a0895599116d
  4. add_vis_pos_emb helper · level 1 · ran · honoured contract · fbc958c1295b5f01
  5. get_chunk_ar_mask helper · level 1 · ran · honoured contract · 3e4ac71a6f12be9d
  6. get_layer_config helper · level 1 · ran · honoured contract · 90a86ebf31046958
  7. suffix_id helper · level 1 · ran · our draft was wrong · d1c561539ef11926
  8. DropPath helper · level 1 · ran · 80161ca59f374d2a
  9. MLP helper · level 0 · unverified · e2256176baee1426
  10. TransformerEncoderLayer model · level 0 · unverified · e730a175de3303f8
  11. get_1d_position_codes helper · level 1 · ran · honoured contract · cbbd048907025ae1
  12. get_2d_position_codes helper · level 1 · ran · honoured contract · 554d30594bb18ba4
  13. get_variable_initializer helper · level 1 · ran · our draft was wrong · 37888d00e9d8c4ab
  14. get_ar_mask helper · level 1 · ran · honoured contract · 508ee551fb33fedc
  15. kronecker_product helper · level 1 · ran · honoured contract · e37890b1d73b280c
  16. FeedForwardLayer helper · level 1 · ran · b9ad5f9347bcd710
  17. positional_encoding helper · level 1 · ran; the contract check could not run (RAISES) · 7a5b750c9c8d5e81
  18. get_angles helper · level 1 · ran · honoured contract · c1a5330036a08dfa

harvested Assembly for Recurrent Interface Networks (RINs) (graph method: Recurrent Interface Networks (RINs)). 16 bricks; individually on a fixture: 13 ran (of which 7 constructed an object rather than computing a result; 11 with no instrument failure: 0 honoured, 3 violated, 8 with no contract checked; 2 where Syntology's instrument failed) · 3 unverified. The assembly has not been run end-to-end.

one of 2 siblings

  1. RIN entry · level 0 · unverified · 5fdaf6caf1863ed1
  2. FeedForward model · level 1 · ran · 1e9eca95c178f7c9
  3. LayerNorm model · level 1 · ran · d070db73849f9743
  4. LearnedSinusoidalPosEmb helper · level 1 · ran · e4a1970252d572fd
  5. RINBlock model · level 0 · unverified · 89c11df753e06cb9
  6. Sequential helper · level 0 · unverified · d6fde31f78c4502f
  7. default helper · level 1 · ran · violated contract · d1ef6b8cb9a28a53
  8. divisible_by helper · level 2 · ran · violated contract · 856d86b39c6005b4
  9. exists helper · level 1 · ran · violated contract · 608e364a9d2376a3
  10. Attention model · level 1 · ran · 305ed1d536ee99b2
  11. LinearAttention helper · level 1 · ran · f8753f538bdf8090
  12. PEG model · level 1 · ran · 2090cfddb3b4ea08
  13. Attend helper · level 1 · ran · df4d0ae2244e374f
  14. MultiHeadedRMSNorm helper · level 1 · ran · 8535fbe5734cf1d3
  15. l1norm helper · level 1 · ran · our draft was wrong · 52649c5bdcbe1d0a
  16. once helper · level 1 · ran · our draft was wrong · 52e10a7d6b0fa036

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="2212.11972") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Image GenerationVideo GenerationVideo Prediction

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Generation ImageNet 128x128 RIN FID 2.75 #8 of 23 Archive leaderboard report
Image Generation ImageNet 128x128 RIN IS 144.1 #8 of 23 Archive leaderboard report
Image Generation ImageNet 256x256 RIN FID 4.51 #85 of 94 Archive leaderboard report
Image Generation ImageNet 64x64 RIN FID 1.23 #8 of 65 Archive leaderboard report
Video Prediction Kinetics-600 12 frames, 64x64 RIN (1000 steps) FVD 10.8 #6 of 16 Archive leaderboard report
Video Prediction Kinetics-600 12 frames, 64x64 RIN (1000 steps) IS 17.7 #6 of 16 Archive leaderboard report
Video Prediction Kinetics-600 12 frames, 64x64 RIN (400 steps) FVD 11.5 #7 of 16 Archive leaderboard report
Video Prediction Kinetics-600 12 frames, 64x64 RIN (400 steps) IS 17.7 #7 of 16 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

Diffusion

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