Papers › Deep Compression Autoencoder for Efficient High-Resolution Diffusion Models

Deep Compression Autoencoder for Efficient High-Resolution Diffusion Models

14 Oct 2024arXiv:2410.10733archive 2025-07-28

Junyu Chen, Han Cai, Junsong Chen, Enze Xie, Shang Yang, Haotian Tang, Muyang Li, Yao Lu, Song Han

We present Deep Compression Autoencoder (DC-AE), a new family of autoencoder models for accelerating high-resolution diffusion models. Existing autoencoder models have demonstrated impressive results at a moderate spatial compression ratio (e.g., 8x), but fail to maintain satisfactory reconstruction accuracy for high spatial compression ratios (e.g., 64x). We address this challenge by introducing two key techniques: (1) Residual Autoencoding, where we design our models to learn residuals based on the space-to-channel transformed features to alleviate the optimization difficulty of high spatial-compression autoencoders; (2) Decoupled High-Resolution Adaptation, an efficient decoupled three-phases training strategy for mitigating the generalization penalty of high spatial-compression autoencoders. With these designs, we improve the autoencoder's spatial compression ratio up to 128 while maintaining the reconstruction quality. Applying our DC-AE to latent diffusion models, we achieve significant speedup without accuracy drop. For example, on ImageNet 512x512, our DC-AE provides 19.1x inference speedup and 17.9x training speedup on H100 GPU for UViT-H while achieving a better FID, compared with the widely used SD-VAE-f8 autoencoder. Our code is available at https://github.com/mit-han-lab/efficientvit.

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.

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

Code

Syntology 30 code samples harvested from 1 repository linked to this paper: 7 ran (of which 6 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 0 where Syntology's instrument failed) · 23 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 30 samples from 1 repository, 7 ran (7 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.

mit-han-lab/efficientvit officialmentioned in papermentioned on GitHubpytorchSyntology: 30 samples, 7 ran (0 instrument-failed) · 23 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

30 samples harvested: 7 ran (of which 6 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 0 where Syntology's instrument failed) · 23 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.

7ran
23unverified

Licence: 0 of the 30 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 mit-han-lab/efficientvit. “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.

mit-han-lab/efficientvit official repository 30 samples, 7 ran (0 instrument-failed) · 23 unverified
ChannelDuplicatingPixelUnshuffleUpSampleLayer efficientvit/models/efficientvit/dc_ae.py ran · metamorphic tier: deterministic fingerprinted Apache-2.0 (permissive) · 5a6b494abcac2208 · report fixture: returned ChannelDuplicatingPixelUnshuffleUpSampleLayer · level recorded 2026-09-22
DCAEConfig efficientvit/models/efficientvit/dc_ae.py ran · metamorphic tier: deterministic Apache-2.0 (permissive) · a7400f7dc57e2876 · report fixture: returned DCAEConfig · level recorded 2026-09-22
DecoderConfig efficientvit/models/efficientvit/dc_ae.py ran · metamorphic tier: deterministic Apache-2.0 (permissive) · 64d381f16614334c · report fixture: returned DecoderConfig · level recorded 2026-09-22
EncoderConfig efficientvit/models/efficientvit/dc_ae.py ran · metamorphic tier: deterministic Apache-2.0 (permissive) · 7c0f5cb15a33304c · report fixture: returned EncoderConfig · level recorded 2026-09-22
IdentityLayer in 7 papers efficientvit/models/efficientvit/dc_ae.py ran · metamorphic tier: invariant fingerprinted Apache-2.0 (permissive) · 7d8080781cf3019c · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22
OpSequential efficientvit/models/efficientvit/dc_ae.py ran · metamorphic tier: deterministic fingerprinted Apache-2.0 (permissive) · a8bff7f7a6484f90 · report fixture: returned OpSequential · level recorded 2026-09-22
PixelUnshuffleChannelAveragingDownSampleLayer efficientvit/models/efficientvit/dc_ae.py ran · metamorphic tier: deterministic Apache-2.0 (permissive) · 447757507c43986b · report fixture: returned PixelUnshuffleChannelAveragingDownSampleLayer · level recorded 2026-09-22
ConvLayer in 26 papers efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · ee46943abe5b1d38 · report
ConvPixelShuffleUpSampleLayer efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · a50b4796b9e7f993 · report
ConvPixelUnshuffleDownSampleLayer efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · c70aaa792533f82b · report
DCAE efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 7b99877bfad9d9db · report
Decoder in 210 papers efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 516380f38de4bf18 · report
EfficientViTBlock efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 644cd11053cbe8fa · report
Encoder in 278 papers efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · e8f172d1d06ebffc · report
GLUMBConv efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 50c192aef38f49ce · report
InterpolateConvUpSampleLayer efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · e3551b0a1b8b4a44 · report
LiteMLA efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · eca2ca74ae04da15 · report
MBConv in 10 papers efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 63542ac85cdf8517 · report
ResidualBlock in 100 papers efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 9d531d6342d2cc9c · report
build_act in 5 papers efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 63bb3de89867ea9d · report
build_block efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · a0bb0bb02b4fd978 · report
build_decoder_project_in_block efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 29728624d884920a · report
build_decoder_project_out_block efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 498fce97a28da008 · report
build_downsample_block efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 0d1e3498712a547c · report
build_encoder_project_in_block efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 1598ee0cf2028172 · report
build_encoder_project_out_block efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · b8d5cdc9329421af · report
build_norm in 5 papers efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 71068d4faddfa7bb · report
build_stage_main efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · c77ad3ad88af1f35 · report
build_upsample_block efficientvit/models/efficientvit/dc_ae.py unverified Apache-2.0 (permissive) · 885f68d08c4d49f1 · 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 Deep Compression Autoencoder (DC-AE) (graph method: Deep Compression Autoencoder (DC-AE)). 30 bricks; individually on a fixture: 7 ran (of which 6 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 0 violated, 7 with no contract checked; 0 where Syntology's instrument failed) · 23 unverified. The assembly has not been run end-to-end.

  1. DCAE entry · level 0 · unverified · 7b99877bfad9d9db
  2. DCAEConfig helper · level 1 · ran · a7400f7dc57e2876
  3. Decoder model · level 0 · unverified · 516380f38de4bf18
  4. Encoder model · level 0 · unverified · e8f172d1d06ebffc
  5. DecoderConfig helper · level 1 · ran · 64d381f16614334c
  6. EncoderConfig helper · level 1 · ran · 7c0f5cb15a33304c
  7. OpSequential helper · level 1 · ran · a8bff7f7a6484f90
  8. build_decoder_project_in_block helper · level 0 · unverified · 29728624d884920a
  9. build_decoder_project_out_block helper · level 0 · unverified · 498fce97a28da008
  10. build_stage_main helper · level 0 · unverified · c77ad3ad88af1f35
  11. build_upsample_block helper · level 0 · unverified · 885f68d08c4d49f1
  12. build_downsample_block helper · level 0 · unverified · 0d1e3498712a547c
  13. build_encoder_project_in_block helper · level 0 · unverified · 1598ee0cf2028172
  14. build_encoder_project_out_block helper · level 0 · unverified · b8d5cdc9329421af
  15. ChannelDuplicatingPixelUnshuffleUpSampleLayer helper · level 1 · ran · 5a6b494abcac2208
  16. ConvLayer helper · level 0 · unverified · ee46943abe5b1d38
  17. ResidualBlock helper · level 0 · unverified · 9d531d6342d2cc9c
  18. build_act helper · level 0 · unverified · 63bb3de89867ea9d
  19. build_norm helper · level 0 · unverified · 71068d4faddfa7bb
  20. build_block helper · level 0 · unverified · a0bb0bb02b4fd978
  21. ConvPixelShuffleUpSampleLayer helper · level 0 · unverified · a50b4796b9e7f993
  22. InterpolateConvUpSampleLayer helper · level 0 · unverified · e3551b0a1b8b4a44
  23. ConvPixelUnshuffleDownSampleLayer helper · level 0 · unverified · c70aaa792533f82b
  24. PixelUnshuffleChannelAveragingDownSampleLayer helper · level 1 · ran · 447757507c43986b
  25. EfficientViTBlock helper · level 0 · unverified · 644cd11053cbe8fa
  26. IdentityLayer helper · level 1 · ran · 7d8080781cf3019c
  27. ResBlock helper · level 0 · unverified · b8654d21821ceafc
  28. GLUMBConv helper · level 0 · unverified · 50c192aef38f49ce
  29. LiteMLA helper · level 0 · unverified · eca2ca74ae04da15
  30. MBConv helper · level 0 · unverified · 63542ac85cdf8517

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="2410.10733") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Image Generation

1 archive task tag without a task page not shown.

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Generation ImageNet 512x512 DC-AE-f32 + USiT-2B FID 1.72 #16 of 52 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