Papers › Perception Encoder: The best visual embeddings are not at the output of the network

Perception Encoder: The best visual embeddings are not at the output of the network

17 Apr 2025arXiv:2504.13181archive 2025-07-28

Daniel Bolya, Po-Yao Huang, Peize Sun, Jang Hyun Cho, Andrea Madotto, Chen Wei, Tengyu Ma, Jiale Zhi, Jathushan Rajasegaran, Hanoona Rasheed, Junke Wang, Marco Monteiro, Hu Xu, Shiyu Dong, Nikhila Ravi, Daniel Li, Piotr Dollár, Christoph Feichtenhofer

We introduce Perception Encoder (PE), a state-of-the-art encoder for image and video understanding trained via simple vision-language learning. Traditionally, vision encoders have relied on a variety of pretraining objectives, each tailored to specific downstream tasks such as classification, captioning, or localization. Surprisingly, after scaling our carefully tuned image pretraining recipe and refining with our robust video data engine, we find that contrastive vision-language training alone can produce strong, general embeddings for all of these downstream tasks. There is only one caveat: these embeddings are hidden within the intermediate layers of the network. To draw them out, we introduce two alignment methods, language alignment for multimodal language modeling, and spatial alignment for dense prediction. Together with the core contrastive checkpoint, our PE family of models achieves state-of-the-art performance on a wide variety of tasks, including zero-shot image and video classification and retrieval; document, image, and video Q&A; and spatial tasks such as detection, depth estimation, and tracking. To foster further research, we are releasing our models, code, and a novel dataset of synthetically and human-annotated videos.

PaperPDFCodeCode Syntology ran

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 18 code samples harvested from 1 repository linked to this paper: 14 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 1 violated, 5 with no contract checked; 8 where Syntology's instrument failed) · 4 unverified. Where Syntology's instrument failed: 4 where our contract draft was wrong; 4 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: community (archive-listed): 18 samples from 1 repository, 14 ran (6 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 between 2026-09-22 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 14 samples that ran has no recorded date.

facebookresearch/perception_models mentioned on GitHubpytorchApache-2.0Syntology: 18 samples, 14 ran (8 instrument-failed) · 4 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

18 samples harvested: 14 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 0 honoured, 1 violated, 5 with no contract checked; 8 where Syntology's instrument failed) · 4 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · violated contract
4ran · our draft was wrong
4ran; the contract check could not run
5ran
4unverified

Licence: 2 of the 18 samples are pointer only, meaning Syntology does not serve that copy's text. For 2 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from facebookresearch/perception_models. “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”. “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.

facebookresearch/perception_models community (archive-listed) 18 samples, 14 ran (8 instrument-failed) · 4 unverified
broadcat in 21 papers apps/detection/DETA_pe/models/pev1.py ran Apache-2.0 (permissive) · 00e3006fbfc5835f · report fixture: signature tensors:list/1/Tensor, dim:int/0/int · returned Tensor · level recorded 2026-09-25
dice_loss in 193 papers
census: 10 implementations from 16 papers, 7 distinct outputs across 4 buckets, one shared input per bucket this sample: bucket 1 of 4, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind float, dtype float32
apps/detection/DETA_pe/models/segmentation.py ran · violated contract fingerprinted Apache-2.0 (permissive) · ac8fe530cdad4d8c · report fixture: signature inputs:Tensor/2/float, targets:Tensor/2/float, num_boxes:int/0/int · returned Tensor · level recorded 2026-09-22
drop_path in 325 papers
census: 27 implementations from 61 papers, 5 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 1 of 2, arg 1: rank 4, kind float, dtype float32
apps/detection/DETA_pe/models/swin.py ran; the contract check could not run (RAISES) Apache-2.0 (permissive) · 87577b3ff9d32712 · report fixture: signature x:Tensor/4/float, drop_prob:float/0/float, training:bool/0/bool, scale_by_keep:bool/0/bool · returned Tensor · level recorded 2026-09-22
get_fed_loss_inds in 12 papers apps/detection/DETA_pe/models/utils_fed_loss.py ran Apache-2.0 (permissive) · bb583e07209fe58b · report fixture: signature gt_classes:Tensor/1/int, num_sample_cats:int/0/int, C:int/0/int, weight:Tensor/1/float · returned Tensor · level recorded 2026-09-24
get_rel_pos in 54 papers
census: 3 implementations from 6 papers, 3 distinct outputs on one shared input
apps/detection/DETA_pe/models/utils_d2.py ran Apache-2.0 (permissive) · edb9869f131321d2 · report fixture: signature q_size:int/0/int, k_size:int/0/int, rel_pos:Tensor/2/float · returned Tensor · level recorded 2026-09-24
nonzero_tuple apps/detection/DETA_pe/models/assigner.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · d28e238107e0aa1b · report fixture: signature x:Tensor/2/float · returned tuple · level recorded 2026-09-24
rotate_half in 424 papers
census: 24 implementations from 96 papers, 4 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 1 of 2, arg 1: rank 2, kind float, dtype float32
apps/detection/DETA_pe/models/pev1.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 58823d9435a8751b · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22
sample_topk_per_gt apps/detection/DETA_pe/models/assigner.py ran Apache-2.0 (permissive) · 626753a2de4f5009 · report fixture: signature pr_inds:Tensor/1/int, gt_inds:Tensor/1/int, iou:Tensor/2/float, k:int/0/int · returned tuple · level recorded 2026-09-25
sigmoid_focal_loss in 122 papers
census: 3 implementations from 9 papers, 2 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 1 of 2, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind float, dtype float32
apps/detection/DETA_pe/models/segmentation.py ran · our draft was wrong fingerprinted Apache-2.0 (permissive) · 5c0711aada67957e · report fixture: signature inputs:Tensor/2/float, targets:Tensor/2/float, num_boxes:int/0/int, alpha:float/0/float, gamma:float/0/float · returned Tensor · level recorded, date not recorded
subsample_labels apps/detection/DETA_pe/models/assigner.py ran Apache-2.0 (permissive) · 19a04bf3d341cedb · report fixture: signature labels:Tensor/1/int, num_samples:int/0/int, positive_fraction:float/0/float, bg_label:int/0/int · returned tuple · level recorded 2026-09-24
window_partition in 377 papers
census: 13 implementations from 17 papers, 5 distinct outputs across 3 buckets, one shared input per bucket
apps/detection/DETA_pe/models/swin.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted Apache-2.0 (permissive) · f9fd6241d935f07b · report fixture: signature x:Tensor/4/float, window_size:int/0/int · returned Tensor · level recorded 2026-09-22
window_partition in 377 papers
census: 13 implementations from 17 papers, 5 distinct outputs across 3 buckets, one shared input per bucket
apps/detection/DETA_pe/models/utils_d2.py ran; the contract check could not run (RAISES) fingerprinted Apache-2.0 (permissive) here; the stored copy is pointer only · 750facb8c52e8864 · report fixture: signature x:Tensor/4/float, window_size:int/0/int · returned tuple · level recorded 2026-09-24
window_reverse in 296 papers apps/detection/DETA_pe/models/swin.py ran · our draft was wrong Apache-2.0 (permissive) · fb32094c6dbece71 · report fixture: signature windows:Tensor/4/float, window_size:int/0/int, H:int/0/int, W:int/0/int · returned Tensor · level recorded 2026-09-22
window_unpartition in 56 papers apps/detection/DETA_pe/models/utils_d2.py ran; the contract check could not run (RAISES) Apache-2.0 (permissive) here; the stored copy is pointer only · 535942f820542fb9 · report fixture: signature windows:Tensor/4/float, window_size:int/0/int, pad_hw:tuple/1/int, hw:tuple/1/int · returned Tensor · level recorded 2026-09-24
batched_soft_nms apps/detection/DETA_pe/models/utils_softnms.py unverified Apache-2.0 (permissive) · 1cf6c2f532b241f1 · report
get_abs_pos in 15 papers apps/detection/DETA_pe/models/pev1.py unverified Apache-2.0 (permissive) · f733619c7c9a0a40 · report
load_class_freq in 14 papers apps/detection/DETA_pe/models/utils_fed_loss.py unverified Apache-2.0 (permissive) · 0547030fc587fd16 · report
soft_nms apps/detection/DETA_pe/models/utils_softnms.py unverified Apache-2.0 (permissive) · 8b89ea91621de82b · report

Tasks

Depth EstimationLanguage ModelingLanguage ModellingObject DetectionVideo ClassificationVideo Understanding

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Object Detection COCO minival PE_spatial (DETA) Params (M) 1900 #1 of 220 Archive leaderboard report
Object Detection COCO minival PE_spatial (DETA) box AP 66.0 #1 of 220 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.

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