Pith. sign in

REVIEW 1 major objections 6 minor 1 cited by

Test-time Alignment-Enhanced Adapter for Vision-Language Models

T0 review · 1 major / 6 minor · reviewed 2026-08-12 · deepseek-v4-flash

Pith's one-line read This paper claims that adjusting text features at test time, via a lightweight gated attention adapter plus a negative cache, beats prior test-time adaptation on OOD and cross-domain benchmarks in 18 minutes on ImageNet.

desk verdict A competent incremental TTA method whose headline gains over TDA look plausible but are not yet statistically secured, mainly due to missing variance estimates and hyperparameters chosen on the eval benchmarks. read the letter →

arxiv 2411.15735 v1 pith:NJQN6C6X submitted 2024-11-24 cs.CV

classification cs.CV
keywords test-timeadaptationvision-languagemodelsCLIPgatedattentionadapterpseudo-labelsnegativecachetextfeaturealignmentdistributionshift
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Test-time adaptation for CLIP-style vision-language models has so far worked by adjusting classification logits, the scores that compare a test image with fixed text prompts. This paper argues that keeping the text features frozen is a bottleneck, and proposes TAEA: a lightweight gated attention adapter trained on the fly to rewrite the text features so they align better with the test distribution, plus a negative cache that guards against noisy pseudo-labels. The claim is that this combination beats the leading test-time adaptation method TDA by 0.75 to 2.66 accuracy points on standard OOD and cross-domain benchmarks, while keeping test time at 18 minutes on ImageNet, two orders of magnitude faster than prompt-tuning methods such as TPT and DiffTPT. If true, it points to text-side adaptation as a cheap and effective complement to logit adjustment.

What carries the argument

The load-bearing object is the gated attention adapter. It is a single-head attention block (implemented with MLP projections as in Eq. 3) that computes $\hat{F} = F^\top \sigma((\omega W_W^\top)(F W_F^\top)^\top/\sqrt{D})$, retrieving the parts of test image features most relevant to each text category; a one-layer MLP gate $f(\cdot)$ produces a scalar that controls how much of $\hat{F}$ is added to the original text features, giving $\hat{\omega}$. Training happens once, after the first $\lambda N$ samples, using one-hot pseudo-labels from the lowest-entropy predictions. The enhancement module is the negative cache from TDA, which contributes $P_{\text{neg}}$ to counteract overconfident or biased predictions from training on pseudo-labels. The theoretical motivation is that the gated attention behaves like a non-local filter, ignoring outlier samples and attending to samples related to the category description.

What would settle it

Take the lowest-entropy quarter of a test set, train the identical adapter using (a) the original pseudo-labels, (b) pseudo-labels with a fixed fraction (say 20%) randomly flipped, and (c) labels drawn uniformly at random; if the accuracy gain over TDA survives (b) and (c) roughly unchanged, the gain does not come from pseudo-label quality, while if it collapses in (c), the low-entropy selection is materially important. A second check is to compare against a version trained on the same samples with ground-truth labels, which upper-bounds the achievable gain.

Watch

Extended reading notes

Core claim

The central claim is that adjusting text features at test time, not just classification logits, is what unlocks better test-time adaptation for vision-language models, and that a small trained adapter can do this without the hours of prompt tuning that earlier text-side methods needed. TAEA's adapter uses the original text embeddings $\omega$ as query and the test image features $F$ as key and value in a single-head attention block implemented with MLPs, producing image features $\hat{F}$ related to each category; a learnable gate $f(\cdot)$ blends these with the original text features to give $\hat{\omega} = \omega + f(\omega) \odot \hat{F}$. The adapter is trained by cross-entropy on pseudo-labels taken from the lowest-entropy CLIP predictions over the first $\lambda N$ test samples ($\lambda = 0.25$). The final prediction adds the adapter's cosine similarity and the TDA negative-cache term to the original CLIP score: $P_{\text{TAEA}} = P_{\text{clip}} + P_{\text{adapter}} + P_{\text{neg}}$. Reported results: average OOD accuracy rises from 65.01 to 65.76 with ViT-B/16 and from 49.58 to 50.13 with ResNet-50 relative to TDA; cross-domain averages rise from 67.53 to 70.03 with ViT-B/16 and from 61.03 to 63.69 with ResNet-50.

Load-bearing premise

The method depends on the low-entropy CLIP predictions from the first quarter of the test set being trustworthy enough to serve as training labels; if those pseudo-labels are wrong in a systematic way, the adapter learns to reinforce CLIP's own biases rather than correct them.

Editorial extensions

If this is right

  • Text features can be adapted at test time with a few minutes of training, making text-side adaptation competitive with logit-side methods on cost.
  • The negative cache of TDA composes additively with the adapter, so future test-time methods can treat logit-side corrections and text-side corrections as additive.
  • The method sets new average accuracy on the ten-dataset cross-domain benchmark with both ResNet-50 and ViT-B/16 backbones, with the largest gains on EuroSAT and DTD.
  • Test-time cost on ImageNet drops to 18 minutes with a single GPU, versus 12h50m for TPT and 34h45m for DiffTPT, narrowing the efficiency gap to the training-free TDA baseline.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The larger cross-domain gain (2.5 points) compared with the OOD gain (0.75 point) suggests the adapter is most valuable when the distribution shift changes what a correct text description looks like, rather than when images merely become harder versions of the same classes.
  • Because the adapter is trained once on the first quarter of the test set, the method is not truly online; a continual variant that updates the adapter as more test samples arrive (for example, with a replay buffer) is a natural extension the paper does not explore.
  • The low-entropy selection threshold and the gating ratio $\gamma$ are hyperparameters that could be benchmark-specific; a per-dataset sensitivity analysis would clarify whether the reported gains depend on careful tuning, which the paper only partially addresses with the $\gamma$ ablation on the OOD benchmark.
  • A testable alternative to raw CLIP entropy for selecting training samples is consistency across augmented views, as used in prompt-tuning methods; the paper does not compare selection strategies.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

1 major / 6 minor

Summary. The paper proposes TAEA, a test-time adaptation method for vision-language models. It trains a lightweight gated attention adapter on low-entropy test samples to adjust CLIP's text features, and combines this adapter with the negative cache from TDA. Experiments on the OOD benchmark (ImageNet and four shifted variants) and the cross-domain benchmark (ten datasets) using ResNet-50 and ViT-B/16 report average accuracy improvements over TDA of 0.55/0.75 percentage points on OOD and 2.66/2.5 on cross-domain, with an 18-minute test time on ImageNet. The paper's central claim is that this simple adapter improves text-to-image alignment at test time more effectively than prior test-time adaptation methods.

Significance. If the reported gains are reproducible, TAEA is a useful and efficient contribution to test-time adaptation for VLMs: it avoids the large training overhead of prompt-tuning methods while modifying text features, and the gated-attention formulation is plausible and clearly described. The paper also provides a directionally consistent ablation and a concrete efficiency comparison. However, the evidence is entirely empirical and the headline margins are small, with no variance estimates, hyperparameters selected on the evaluation benchmarks, and baselines copied from the TDA paper rather than re-run. These issues make the central claim fragile as it stands, though they are addressable with additional experiments.

major comments (1)
  1. [Section IV-A (Baselines)] The statement 'All results of the methods compared in the table are obtained from the [7] paper' means that TDA, TPT, and DiffTPT numbers are copied rather than re-run under identical conditions, including the same preprocessing, test-time protocol, and evaluation code. Since the claimed contribution is a small average improvement over TDA, a direct re-implementation under the same harness is needed to rule out differences in evaluation details. At minimum, please re-run TDA using its released code and report the resulting numbers.
minor comments (6)
  1. [Equation (3)] The dimension notation in Eq. (3) is not fully specified: please define the shapes of W_W and W_F and state over which axis the softmax is applied, so that the expression can be verified.
  2. [Section IV-A, first sentence] The citation 'Consistent with prior works [6], [19]' is problematic: [19] is Loshchilov and Hutter's AdamW reference, not a prior test-time adaptation work. Please correct the citation to the intended prior benchmark papers.
  3. [Figure 2] The captions for Figure 2 are inconsistent ('Imagenet' should be 'ImageNet') and the subfigures lack axis labels and legends, which makes the ablation results hard to read.
  4. [Section III-B, last paragraph] The sentence 'The effectiveness of adjusting text features is theoretically guaranteed by non-local filters [16]-[18]' is too strong and vague; the cited works do not provide a formal guarantee for this adaptation setting. Please rephrase this as an intuition or provide a precise statement of the guarantee.
  5. [Section IV-A (Implementation Details)] The paper says that all database-related hyperparameters are consistent with TDA, but it does not list the cache sizes, queue lengths, negative-cache coefficients, or the exact selection criterion for low-entropy samples. Please provide these details for reproducibility.
  6. [Abstract / reproducibility] The GitHub URL appears only in the abstract; please ensure the code repository is public and link it in the paper body, and note the PyTorch version and evaluation harness used for the timing measurements in Table III.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: TAEA's adapter is trained on CLIP pseudo-labels but evaluated on ground-truth labels; the reported gains are empirical, not definitional.

full rationale

The paper's derivation chain is a standard test-time adaptation pipeline rather than a circular reduction. The adapter module (Eqs. 3-5) is a gated attention block inspired by Meta-Adapter [12] and the enhancement module is TDA's negative cache [7]; neither component is defined in terms of the target accuracy. The pseudo-labels in Section III-B are generated from low-entropy CLIP predictions, but the reported accuracy is measured against ground-truth labels on ImageNet, OOD, and cross-domain benchmarks, so the result is not self-confirming by construction. The choice of gamma = 0.6 in Fig. 2(b) using the OOD benchmark is a hyperparameter-selection concern that could inflate the reported margin, but it is not a circular derivation: the accuracy numbers are not algebraically determined by the chosen gamma, and the comparison to TDA is based on independently published baselines. Baseline numbers are copied from TDA [7] rather than re-run, which is an experimental rigor issue rather than circularity. No load-bearing self-citations appear; citations [7] and [12] are to external research groups. The central claim therefore has independent empirical content and does not reduce to its inputs by construction.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The method introduces no new theoretical entities, but it depends on pseudo-label reliability, a borrowed attention mechanism, and inherited TDA cache settings. The two explicit hyperparameters, gamma and lambda, are fitted on the same benchmarks used for evaluation.

free parameters (5)
  • gamma = 0.6
    Weight controlling adapter logit contribution in Eq. (5); selected via ablation on the OOD benchmark in Fig. 2(b), not on a held-out set.
  • lambda = 0.25
    Fraction of test set observed before training the adapter; chosen to align with TDA timing but not independently justified or ablated.
  • adapter learning rate = 0.001
    AdamW learning rate for the gated attention block, chosen as a default; no sensitivity analysis is reported.
  • adapter epochs and batch size = 3 epochs, batch size 3
    Training schedule for the adapter module in Section IV-A; no ablation is provided for these choices.
  • TDA negative cache hyperparameters = not reported
    The paper says all database-related hyperparameters are consistent with TDA but does not state queue sizes or logit scaling, so the borrowed module's settings are not independently auditable.
assumptions (4)
  • domain assumption Low-entropy test samples produce reliable pseudo-labels for training the adapter.
    Used in Section III-B to select the first lambda*N samples and generate one-hot pseudo-labels; if wrong, the adapter trains on mislabeled data and the reported gains could be spurious.
  • ad hoc to paper A gated attention adapter can reduce the text-image distribution gap at test time.
    The core mechanism in Eq. (3)-(4) is taken from Meta-Adapter and applied at test time; the paper cites non-local filters as theoretical support but provides no proof for the test-time setting.
  • domain assumption Adapting text features while keeping the visual encoder frozen is sufficient to recover alignment under distribution shift.
    The method only updates text features in Eq. (4) and never adapts the image encoder; performance depends on the premise that most distribution shift is in the text-image alignment direction.
  • ad hoc to paper The negative cache from TDA remains beneficial when combined with the adapter.
    The enhancement module is taken from TDA; no independent analysis explains why the two modules interact positively rather than double-counting the same logits.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Test-time Alignment-Enhanced Adapter for Vision-Language Models." pith.science (2026). https://pith.science/paper/NJQN6C6X

@misc{pith2026241115735,
  author       = {Pith},
  title        = {Pith review of: Test-time Alignment-Enhanced Adapter for Vision-Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NJQN6C6X}},
  note         = {Machine review of arXiv:2411.15735}
}
read the original abstract

Test-time adaptation with pre-trained vision-language models (VLMs) has attracted increasing attention for tackling the issue of distribution shift during the test phase. While prior methods have shown effectiveness in addressing distribution shift by adjusting classification logits, they are not optimal due to keeping text features unchanged. To address this issue, we introduce a new approach called Test-time Alignment-Enhanced Adapter (TAEA), which trains an adapter with test samples to adjust text features during the test phase. We can enhance the text-to-image alignment prediction by utilizing an adapter to adapt text features. Furthermore, we also propose to adopt the negative cache from TDA as enhancement module, which further improves the performance of TAEA. Our approach outperforms the state-of-the-art TTA method of pre-trained VLMs by an average of 0.75% on the out-of-distribution benchmark and 2.5% on the cross-domain benchmark, with an acceptable training time. Code will be available at https://github.com/BaoshunWq/clip-TAEA.

Figures

Figures reproduced from arXiv: 2411.15735 by the authors.

Figure 1
Figure 1. Overall pipeline of the proposed method. The adapter module aims to [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The results of ablation study. tures, improving all accuracy by 6.03% and 4.93% on average, respectively. Furthermore, compared to the current state-of￾the-art method TDA, our method outperforms TDA on both ResNet-50 and ViT-B/16 architectures, improving all accuracy by 2.66% and 2.5% on average, respectively. C. Ablation Study To validate the effectiveness of our proposed method, we conducted ablation experiments o… view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Adapting Vision-Language Models Without Labels: A Comprehensive Survey

    cs.LG 2025-08 conditional novelty 5.0 of 10

    A survey that organizes unsupervised vision-language model adaptation by unlabeled-data availability into four paradigms: data-free transfer, domain transfer, episodic test-time, and online test-time adaptation.

Reference graph

Works this paper leans on

38 extracted references · 5 canonical work pages · cited by 1 Pith paper

  1. [7]

    Efficient test-time adaptation of vision-language models,

    A. Karmanov, D. Guan, S. Lu, A. El Saddik, and E. Xing, “Efficient test-time adaptation of vision-language models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024, pp. 14 162–14 171

  2. [1]

    Tent: Fully test-time adaptation by entropy minimization,

    D. Wang, E. Shelhamer, S. Liu, B. Olshausen, and T. Darrell, “Tent: Fully test-time adaptation by entropy minimization,” in International Conference on Learning Representations , 2020

  3. [2]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark et al., “Learning transferable visual models from natural language supervision,” in International conference on machine learning . PMLR, 2021, pp. 8748–8763

  4. [3]

    Delving into out-of- distribution detection with vision-language representations,

    Y . Ming, Z. Cai, J. Gu, Y . Sun, W. Li, and Y . Li, “Delving into out-of- distribution detection with vision-language representations,” Advances in neural information processing systems , vol. 35, pp. 35 087–35 102, 2022

  5. [4]

    Zegclip: Towards adapting clip for zero-shot semantic segmentation,

    Z. Zhou, Y . Lei, B. Zhang, L. Liu, and Y . Liu, “Zegclip: Towards adapting clip for zero-shot semantic segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 11 175–11 185

  6. [5]

    Test-time prompt tuning for zero-shot generalization in vision-language models,

    M. Shu, W. Nie, D.-A. Huang, Z. Yu, T. Goldstein, A. Anandkumar, and C. Xiao, “Test-time prompt tuning for zero-shot generalization in vision-language models,” Advances in Neural Information Processing Systems, vol. 35, pp. 14 274–14 289, 2022

  7. [6]

    Diverse data augmentation with diffusions for effective test-time prompt tuning,

    C.-M. Feng, K. Yu, Y . Liu, S. Khan, and W. Zuo, “Diverse data augmentation with diffusions for effective test-time prompt tuning,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 2704–2714

  8. [8]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” Advances in neural information processing systems , vol. 30, 2017

Show all 38 references
  1. [9]

    Reducing domain gap by reducing style bias,

    H. Nam, H. Lee, J. Park, W. Yoon, and D. Yoo, “Reducing domain gap by reducing style bias,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 8690–8699

  2. [10]

    Clip-adapter: Better vision-language models with feature adapters,

    P. Gao, S. Geng, R. Zhang, T. Ma, R. Fang, Y . Zhang, H. Li, and Y . Qiao, “Clip-adapter: Better vision-language models with feature adapters,” International Journal of Computer Vision , vol. 132, no. 2, pp. 581–595, 2024

  3. [11]

    Tip-adapter: Training-free adaption of clip for few-shot classification,

    R. Zhang, W. Zhang, R. Fang, P. Gao, K. Li, J. Dai, Y . Qiao, and H. Li, “Tip-adapter: Training-free adaption of clip for few-shot classification,” in European conference on computer vision . Springer, 2022, pp. 493– 510

  4. [12]

    Meta-adapter: An online few-shot learner for vision-language model,

    L. Song, R. Xue, H. Wang, H. Sun, Y . Ge, Y . Shanet al., “Meta-adapter: An online few-shot learner for vision-language model,” Advances in Neural Information Processing Systems , vol. 36, pp. 55 361–55 374, 2023

  5. [13]

    Sam-adapter: Adapting segment anything in underperformed scenes,

    T. Chen, L. Zhu, C. Deng, R. Cao, Y . Wang, S. Zhang, Z. Li, L. Sun, Y . Zang, and P. Mao, “Sam-adapter: Adapting segment anything in underperformed scenes,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops , October 2023, pp. 3367–3375

  6. [14]

    T2i- adapter: Learning adapters to dig out more controllable ability for text- to-image diffusion models,

    C. Mou, X. Wang, L. Xie, Y . Wu, J. Zhang, Z. Qi, and Y . Shan, “T2i- adapter: Learning adapters to dig out more controllable ability for text- to-image diffusion models,” in Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 5, 2024, pp. 4296–4304

  7. [15]

    On the effectiveness of adapter-based tuning for pretrained language model adaptation,

    R. He, L. Liu, H. Ye, Q. Tan, B. Ding, L. Cheng, J. Low, L. Bing, and L. Si, “On the effectiveness of adapter-based tuning for pretrained language model adaptation,” in Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th Interna...

  8. [16]

    Non-local neural net- works,

    X. Wang, R. Girshick, A. Gupta, and K. He, “Non-local neural net- works,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 7794–7803

  9. [17]

    Deep hdr imaging via a non-local network,

    Q. Yan, L. Zhang, Y . Liu, Y . Zhu, J. Sun, Q. Shi, and Y . Zhang, “Deep hdr imaging via a non-local network,” IEEE Transactions on Image Processing, vol. 29, pp. 4308–4322, 2020

  10. [18]

    Gcnet: Non-local networks meet squeeze-excitation networks and beyond,

    Y . Cao, J. Xu, S. Lin, F. Wei, and H. Hu, “Gcnet: Non-local networks meet squeeze-excitation networks and beyond,” in Proceedings of the IEEE/CVF international conference on computer vision workshops , 2019, pp. 0–0

  11. [19]

    Decoupled weight decay regularization,

    I. Loshchilov and F. Hutter, “Decoupled weight decay regularization,” in International Conference on Learning Representations , 2017

  12. [20]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in 2009 IEEE conference on computer vision and pattern recognition . Ieee, 2009, pp. 248–255

  13. [21]

    Natural adversarial examples,

    D. Hendrycks, K. Zhao, S. Basart, J. Steinhardt, and D. Song, “Natural adversarial examples,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , 2021, pp. 15 262–15 271

  14. [22]

    Do imagenet clas- sifiers generalize to imagenet?

    B. Recht, R. Roelofs, L. Schmidt, and V . Shankar, “Do imagenet clas- sifiers generalize to imagenet?” in International conference on machine learning. PMLR, 2019, pp. 5389–5400

  15. [23]

    The many faces of robustness: A critical analysis of out-of-distribution generalization,

    D. Hendrycks, S. Basart, N. Mu, S. Kadavath, F. Wang, E. Dorundo, R. Desai, T. Zhu, S. Parajuli, M. Guo et al. , “The many faces of robustness: A critical analysis of out-of-distribution generalization,” in Proceedings of the IEEE/CVF international conference on computer visio...

  16. [24]

    Learning robust global representations by penalizing local predictive power,

    H. Wang, S. Ge, Z. Lipton, and E. P. Xing, “Learning robust global representations by penalizing local predictive power,” Advances in Neural Information Processing Systems , vol. 32, 2019

  17. [25]

    Cats and dogs,

    O. M. Parkhi, A. Vedaldi, A. Zisserman, and C. Jawahar, “Cats and dogs,” in 2012 IEEE conference on computer vision and pattern recog- nition. IEEE, 2012, pp. 3498–3505

  18. [26]

    Eurosat: A novel dataset and deep learning benchmark for land use and land cover classi- fication,

    P. Helber, B. Bischke, A. Dengel, and D. Borth, “Eurosat: A novel dataset and deep learning benchmark for land use and land cover classi- fication,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , vol. 12, no. 7, pp. 2217–2226, 2019

  19. [27]

    Fine- grained visual classification of aircraft,

    S. Maji, E. Rahtu, J. Kannala, M. Blaschko, and A. Vedaldi, “Fine- grained visual classification of aircraft,” arXiv preprint arXiv:1306.5151, 2013

  20. [28]

    Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories,

    L. Fei-Fei, R. Fergus, and P. Perona, “Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories,” in 2004 conference on computer vision and pattern recognition workshop . IEEE, 2004, pp. 178–178

  21. [29]

    Ucf101: A dataset of 101 human actions classes from videos in the wild,

    K. Soomro, A. R. Zamir, and M. Shah, “Ucf101: A dataset of 101 human actions classes from videos in the wild,”arXiv preprint arXiv:1212.0402, 2012

  22. [30]

    3d object representations for fine-grained categorization,

    J. Krause, M. Stark, J. Deng, and L. Fei-Fei, “3d object representations for fine-grained categorization,” in Proceedings of the IEEE interna- tional conference on computer vision workshops , 2013, pp. 554–561

  23. [31]

    Describing textures in the wild,

    M. Cimpoi, S. Maji, I. Kokkinos, S. Mohamed, and A. Vedaldi, “Describing textures in the wild,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2014, pp. 3606–3613

  24. [32]

    Automated flower classification over a large number of classes,

    M.-E. Nilsback and A. Zisserman, “Automated flower classification over a large number of classes,” in 2008 Sixth Indian conference on computer vision, graphics & image processing . IEEE, 2008, pp. 722–729

  25. [33]

    Food-101–mining discriminative components with random forests,

    L. Bossard, M. Guillaumin, and L. Van Gool, “Food-101–mining discriminative components with random forests,” in Computer vision– ECCV 2014: 13th European conference, zurich, Switzerland, September 6-12, 2014, proceedings, part VI 13 . Springer, 2014, pp. 446–461

  26. [34]

    Sun database: Large-scale scene recognition from abbey to zoo,

    J. Xiao, J. Hays, K. A. Ehinger, A. Oliva, and A. Torralba, “Sun database: Large-scale scene recognition from abbey to zoo,” in 2010 IEEE computer society conference on computer vision and pattern recognition. IEEE, 2010, pp. 3485–3492

  27. [35]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition , 2016, pp. 770–778

  28. [36]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” arXiv preprint arXiv:2010.11929 , 2020

  29. [37]

    Learning to prompt for vision- language models,

    K. Zhou, J. Yang, C. C. Loy, and Z. Liu, “Learning to prompt for vision- language models,” International Journal of Computer Vision , vol. 130, no. 9, pp. 2337–2348, 2022

  30. [38]

    Conditional prompt learning for vision-language models,

    ——, “Conditional prompt learning for vision-language models,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 16 816–16 825

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.