Pith. sign in

REVIEW 19 cited by

Vision Transformer Adapter for Dense Predictions

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2205.08534 v4 pith:Z2D2QS46 submitted 2022-05-17 cs.CV

Zhe Chen , Yuchen Duan , Wenhai Wang , Junjun He , Tong Lu , Jifeng Dai , Yu Qiao This is my paper
classification cs.CV
keywords densevit-adapteradapterplaintasksvision-specificbiasesdata
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

This work investigates a simple yet powerful dense prediction task adapter for Vision Transformer (ViT). Unlike recently advanced variants that incorporate vision-specific inductive biases into their architectures, the plain ViT suffers inferior performance on dense predictions due to weak prior assumptions. To address this issue, we propose the ViT-Adapter, which allows plain ViT to achieve comparable performance to vision-specific transformers. Specifically, the backbone in our framework is a plain ViT that can learn powerful representations from large-scale multi-modal data. When transferring to downstream tasks, a pre-training-free adapter is used to introduce the image-related inductive biases into the model, making it suitable for these tasks. We verify ViT-Adapter on multiple dense prediction tasks, including object detection, instance segmentation, and semantic segmentation. Notably, without using extra detection data, our ViT-Adapter-L yields state-of-the-art 60.9 box AP and 53.0 mask AP on COCO test-dev. We hope that the ViT-Adapter could serve as an alternative for vision-specific transformers and facilitate future research. The code and models will be released at https://github.com/czczup/ViT-Adapter.

Discussion (0). Sign in to comment.

Forward citations

Cited by 19 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 204 citations worldwide. Full citation record

  1. LiLa-WAM: Lightweight Latent Reasoning World-Action Model for Robotic Manipulation

    cs.RO 2026-08 conditional novelty 6.0 of 10

    A 0.5B parameter latent world-action model trains end-to-end on a single GPU and reaches 90.48% average success on 50 RoboTwin 2.0 tasks with a language-free Visual Transition Token for task specification.

  2. From Multi-Resolution Cells to Gigapixel Whole Slide Images Foundation Model for Computational Pathology

    cs.CV 2026-08 conditional novelty 6.0 of 10

    MRPT, a multi-resolution hierarchical transformer pre-trained on 36K whole-slide images, is reported to outperform prior pathology foundation models on 34 classification, captioning, and VQA datasets.

  3. iFAN: Inference-Aware Learning for Plain Mask Transformers

    cs.CV 2026-08 conditional novelty 6.0 of 10

    iFAN improves query-based mask transformers by training a mask-quality head for score ranking and distilling better intermediate-layer predictions into the final layer, yielding consistent gains at no inference-time cost.

  4. FBLayout: Optimizing Memory Layout for Efficient LLM Finetuning on Mobile GPUs

    cs.AI 2026-07 conditional novelty 6.0 of 10

    A tile-based memory layout for mobile GPUs that unifies forward and backward data access, eliminating most transpose/reshape overhead and speeding LLM fine-tuning 2.2–5.7× in the paper's measurements.

  5. REAL-OW: Rehearsal-free Open World Object Detection with Low-Rank Adaptation and Dual-Stage Objectness Modeling

    cs.CV 2026-07 conditional novelty 6.0 of 10

    A rehearsal-free open-world detector using collaborative LoRA adapters and dual-stage objectness modeling outperforms exemplar-replay OWOD methods on standard benchmarks.

  6. UNITY: Attention Flow Networks for Adaptive Conditioning in Diffusion

    cs.CV 2026-06 unverdicted novelty 6.0 of 10

    UNITY is a two-stage adapter with Morphable Attention Flow networks for efficient single and composite conditioning in diffusion-based image generation.

  7. Foundation Model-Driven Semantic Change Detection in Remote Sensing Imagery

    cs.CV 2026-02 unverdicted novelty 6.0 of 10

    PerASCD sets new state-of-the-art Sek scores on SECOND and LandsatSCD datasets by using a modular cascaded gated decoder on PerA foundation model features plus a new consistency loss.

  8. Exploring the Rashomon Set for Concept-Based Models

    cs.LG 2025-11 conditional novelty 6.0 of 10

    A shared frozen backbone plus per-model LoRA adapters and a concept-diversity loss trains a set of accurate CBMs that reason through different concepts.

  9. Live(r) Die: Predicting Survival in Colorectal Liver Metastasis

    cs.CV 2025-09 conditional novelty 6.0 of 10

    A fully automated pre/post-contrast MRI framework, combining prompt-based segmentation with autoencoder multiple-instance survival analysis, improves CRLM post-surgery survival prediction over clinical and genomic bio...

  10. FoMo4Wheat: Toward reliable crop vision foundation models with globally curated data

    cs.CV 2025-09 conditional novelty 6.0 of 10

    Training a vision transformer on 2.5 million wheat images outperforms general-domain backbones across ten crop vision tasks.

  11. MPT: Motion Prompt Tuning for Micro-Expression Recognition

    cs.CV 2025-08 unverdicted novelty 6.0 of 10

    Motion Prompt Tuning with motion magnification and Gaussian tokenization claims state-of-the-art micro-expression recognition on three benchmarks.

  12. Latest Object Memory Management for Temporally Consistent Video Instance Segmentation

    cs.CV 2025-07 conditional novelty 6.0 of 10

    LOMM achieves 54.0 AP on YouTube-VIS 2022 (offline) and 48.2 AP online, via foreground-probability-weighted memory and occupancy-guided decoupled association.

  13. Mamba Guided Boundary Prior Matters: A New Perspective for Generalized Polyp Segmentation

    cs.CV 2025-07 conditional novelty 6.0 of 10

    SAM-MaGuP, a SAM-based polyp segmentation model with a 1D-2D Mamba adapter and boundary distillation, reports state-of-the-art mDice/mIoU on five public colonoscopy datasets.

  14. Efficient Medical Vision-Language Alignment Through Adapting Masked Vision Models

    cs.CV 2025-06 conditional novelty 6.0 of 10

    ALTA adapts a frozen masked-pretrained X-ray encoder to language with 8% trainable parameters and temporal-multiview inputs, improving medical retrieval and zero-shot classification.

  15. AI-driven Remote Facial Skin Hydration and TEWL Assessment from Selfie Images: A Systematic Solution

    cs.CV 2025-09 conditional novelty 5.0 of 10

    Selfie images can be used to predict skin hydration and water loss with R2 up to about 0.35, using a new dataset of 336 panelists and an adapter-based vision transformer.

  16. Beyond Appearance: Geometric Cues for Robust Video Instance Segmentation

    cs.CV 2025-07 conditional novelty 5.0 of 10

    Concatenating monocular depth maps as an extra input channel improves video instance segmentation and reaches 56.2 AP, a new state of the art on OVIS.

  17. Colorectal Cancer Tumor Grade Segmentation in Digital Histopathology Images: From Giga to Mini Challenge

    cs.CV 2025-07 conditional novelty 4.0 of 10

    A challenge summary showing top-performing deep learning methods improve colorectal cancer tumor grade segmentation on the METU CCTGS dataset, with the winner reaching 70.2 macro F-score.

  18. SAILViT: Towards Robust and Generalizable Visual Backbones for MLLMs via Gradual Feature Refinement

    cs.CV 2025-07 conditional novelty 4.0 of 10

    A three-stage coarse-to-fine training recipe for vision backbones produces consistent benchmark gains for lightweight multimodal LLMs.

  19. Data-Efficient Challenges in Visual Inductive Priors: A Retrospective

    cs.CV 2025-06 conditional novelty 3.0 of 10

    A retrospective of four data-limited computer vision challenges finds that ensembles and heavy augmentation, not novel inductive priors, drove winning performance.

Pith tools