Pith. sign in

REVIEW 5 major objections 5 minor 33 references

Adaptive Prompt Tuning: Vision Guided Prompt Tuning with Cross-Attention for Fine-Grained Few-Shot Learning

T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read Adaptive Prompt Tuning makes CLIP's text prompts react to the image at hand, and this per-image refinement is what lifts fine-grained few-shot accuracy over static prompt tuning.

desk verdict APT is a plausible cross-attention prompt-tuning variant with promising fine-grained few-shot results, but the paper never validates against CoCoOp properly, so the central claim of beating adaptive prompting is unproven. read the letter →

arxiv 2412.14640 v2 pith:MOAZCU7V submitted 2024-12-19 cs.CV cs.AIcs.LG

classification cs.CVcs.AIcs.LG
keywords CLIPfew-shotlearningfine-grainedimageclassificationprompttuningcross-attentionadaptiveinferenceuncertaintyquantificationMonte-Carlodropout
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

The paper proposes Adaptive Prompt Tuning (APT), a method that makes CLIP's text prompts specific to each test image before classifying it. Instead of learning one fixed prompt (CoOp) or fixed visual tokens (VPT), APT uses a cross-attention layer in which text-encoded prompts act as queries and the image's ViT patch features act as keys and values; only that layer is trained on the few labeled examples while CLIP stays frozen. The claim is that this dynamic, image-guided refinement helps most when classes differ only subtly and vary a lot within a class, as in FGVC Aircraft and CUBirds, and that adding Monte-Carlo Dropout yields calibrated confidence estimates without sacrificing accuracy. If correct, this offers a parameter-lean way to adapt foundation models to specialist classification tasks with very few labels and trustworthy confidence.

What carries the argument

The load-bearing component is the cross-attention module inserted between CLIP's frozen text encoder and the cosine-similarity classifier. Queries come from the text-feature matrix W (one column per class prompt), keys and values from the sequence of image patch features z, so the refined prompt is conditioned on the specific image; layer normalization, a feed-forward block, skip connections, and 20% dropout complete the module, and the dropout doubles as the stochastic sampling mechanism for MC Dropout. This is the only trained part of the model, so the method adapts per image without unfreezing CLIP or storing the training set at inference.

What would settle it

Run APT against CoCoOp on the same 1, 2, 4, 8, 16-shot protocol on FGVC Aircraft and CUBirds, and ablate the cross-attention layer by replacing it with a static learned prompt of equal parameter count; if APT does not match or beat CoCoOp or removing the layer does not lower Aircraft accuracy, the central claim fails.

Watch

Extended reading notes

Core claim

The central discovery claimed is that a single cross-attention refinement of CLIP text embeddings by image patches is enough to outperform static prompt tuning on fine-grained few-shot classification. APT feeds the text features as queries and the image patch tokens as keys and values, producing tuned text features $W'$ that are then compared with the image CLS token by cosine similarity. In experiments on FGVC Aircraft, Oxford Flowers, and CUBirds, APT reports higher accuracy than CoOp and VPT at most shot counts; on Aircraft it rises from 27% one-shot to 47% at 16 shots versus a 17% zero-shot baseline, and on the base-to-new generalization split its harmonic-mean F1 beats CoOp and CoCoOp on Aircraft and Flowers while matching CLIP on CUBirds. The paper also claims that keeping dropout active at inference (MC Dropout, 100 samples) gives uncertainty estimates that separate correct from incorrect predictions and whose ECE falls as shots increase.

Load-bearing premise

The premise that one thin cross-attention layer, trained on only a handful of examples per class, can generalize to unseen images and re-weight text prompts better than static or image-conditioned prompts; this is assumed in the method definition rather than established by an ablation.

Editorial extensions

If this is right

  • On FGVC Aircraft, APT's accuracy rises from 27% at one shot to 47% at 16 shots, a larger gain over the 17% zero-shot CLIP baseline than CoOp or VPT achieve.
  • On Oxford Flowers, APT reaches 84% with a single example per class and 97% at 16 shots, with CoOp close behind.
  • On the base-to-new generalization split, APT's harmonic-mean F1 exceeds CoOp and CoCoOp on FGVC Aircraft (36.46 vs. 28.75 and 27.74) and Oxford Flowers (83.23 vs. 74.06 and 81.71), and matches the CLIP baseline on CUBirds.
  • Keeping dropout active at inference turns APT into a stochastic model whose expected calibration error falls as the number of training shots grows, giving confidence estimates that track accuracy.

Reading between the lines

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

  • A natural port would be to other frozen dual-encoder models: the cross-attention layer is small and trained independently of CLIP, so the same recipe could be tried on any vision-language model that exposes patch tokens and text embeddings; the paper only demonstrates it with CLIP ViT-B/16 on four datasets.
  • The appendix's variance analysis predicts where dynamic prompting will help: datasets with high intra-class and low inter-class image-feature variance should show the largest APT advantage, so ranking additional fine-grained datasets by that gap is a direct testable extension.
  • The Caltech101 results suggest that cluttered backgrounds can pull the cross-attention toward irrelevant patches; masking salient regions or adding a background-suppression term would test whether that is the cause.
  • MCD at a fixed 20% dropout is one calibration recipe; attaching Deep Ensembles or variational inference to the same cross-attention layer would test whether the residual overconfidence seen in high-confidence bins is a property of dropout or of the layer itself.
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

5 major / 5 minor

Summary. The paper proposes Adaptive Prompt Tuning (APT), a cross-attention module that refines CLIP text-prompt embeddings using image features at inference time, with frozen CLIP encoders and a lightweight trainable layer. The method is evaluated on fine-grained few-shot classification (CUBirds, Oxford Flowers, FGVC Aircraft, and Caltech101) against CoOp and VPT, and is augmented with Monte-Carlo Dropout for uncertainty quantification, reporting ECE, reliability plots, confidence-uncertainty plots, and OOD analysis. The central claims are that APT improves accuracy over static prompt tuning and that MCD yields calibrated uncertainty without sacrificing accuracy.

Significance. If the claims are substantiated, the paper would offer a simple, parameter-efficient way to make prompt tuning image-conditional, which is a plausible extension of CoCoOp-style conditional prompting. The explicit integration of MCD with the prompt-tuning pipeline is a practical direction, and the release of code is a strength. However, the empirical evidence as presented is incomplete: the main few-shot curves omit the closest adaptive-prompt baseline (CoCoOp), the single CoCoOp comparison reports an implausibly low number, the ECE formula is mis-specified, and accuracy results are reported without variance or significance tests. These gaps prevent the paper from establishing its claimed state-of-the-art performance, though they appear fixable within the manuscript's scope.

major comments (5)
  1. [Section 4.1, Table 1] The central claim that APT outperforms existing adaptive prompting is not tested against the closest baseline. Figure 4 compares only with CoOp and VPT, which are static-prompt methods, while APT's premise is image-conditioned adaptation; CoCoOp (Zhou et al., 2022a) is the required baseline. The only CoCoOp results, in Table 1, report CUBirds New-class accuracy 8.04 and F1 14.40, far below published CoCoOp values on the same base-to-new protocol (e.g., roughly 60-70 on new classes in the original CoCoOp paper). This discrepancy indicates a non-standard implementation, so the table does not support the claim that APT improves over conditional prompting. Please add CoCoOp to the main few-shot curves and correct the CoCoOp results.
  2. [Section 3.4, ECE definition] The Expected Calibration Error formula is mis-specified. The weight |B_p|/N_p is always 1 because N_p is defined as the number of data points in bin B_p, reducing ECE to an unweighted average of per-bin gaps rather than the standard sample-weighted ECE. The denominator should be the total number of data points N, or the estimator should be explicitly stated as equal-weight ECE. Additionally, the summation index M in the formula conflicts with the earlier use of M as the MC sample size. Please recompute all ECE values in Section 4.2 with a corrected formula.
  3. [Section 3.2, Figure 1] The method description is ambiguous about what serves as keys and values in the cross-attention module. The text states that 'keys k and values v are created from z' where z is the CLS token, a single d-dimensional vector, which would provide only one key/value and cannot implement meaningful cross-attention over image regions. Figure 1, however, shows the 'Sequence of Patches' entering the cross-attention block. Please clarify whether the keys/values are the patch tokens or the CLS token, and make the text and figure consistent; this is essential for reproducing the method.
  4. [Section 4.1 and Table 1] Accuracy results are reported as averages of 3 runs without standard deviations or significance testing. In Figure 4, several curves are close (e.g., Flowers102 at 8 and 16 shots), and a 3-run average without spread does not demonstrate a reliable difference. Please report per-seed results or add standard deviations/confidence intervals for all datasets and shot values, and for the base/new results in Table 1.
  5. [Sections 1, 3.4, 4.2] The paper claims that Monte-Carlo Dropout improves calibration 'without drops in accuracy,' but the accuracy results in Figure 4 do not state whether inference used MC sampling or a deterministic forward pass, and no direct comparison of accuracy with versus without MCD is reported. Please specify the inference protocol for all accuracy results and provide a head-to-head comparison (deterministic APT vs APT+MCD) to substantiate the 'no drop in accuracy' claim.
minor comments (5)
  1. [Section 3.2] The paragraph beginning 'As shown in the preliminary experiments ...' through '... general architecture of the proposed model.' is duplicated verbatim; please remove the repetition.
  2. [Figure 1 caption] The caption contains the typo 'netter fit the images' which should be 'better fit the images'.
  3. [Section 4.2] The text repeatedly misspells 'FGVC Aircraft' as 'FVGC Aircraft' in the calibration discussion; please correct it.
  4. [Figure 5] Figure 5 appears only as a caption with no corresponding plot in the provided text; please add the figure or remove the reference.
  5. [Section 3.3] The VPT description states that learnable tokens are added 'to the Transformer layer' without specifying whether this applies to one or all layers; the notation X = [P1,...,PK,x1,...,xN] would benefit from a precise definition of layer indices and token dimensions.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: APT accuracy is measured on held-out test data, hyperparameters are fixed in advance, and the load-bearing comparisons are external baselines.

full rationale

The paper's derivation chain is not circular. APT's cross-attention layer is trained on labeled few-shot training images while CLIP encoders are frozen, and the reported accuracies are measured on held-out test data (Figure 4, Table 1) averaged over three seeds. Hyperparameters (dropout 0.2, 50-150 epochs, SGD with lr 0.001) are fixed in Section 3.6 rather than fitted to test accuracy. Conditioning text prompts on the test image patches is a modeling choice, not a definitional identity: the classification probability in Eq. 1 still uses class-specific text features and temperature-scaled cosine similarity, so the prediction is not the training objective by construction. The MCD/UQ analysis reuses the test predictions but evaluates calibration (ECE, reliability plots) rather than presenting the calibration metric as an independent prediction; the absence of a deterministic non-MCD baseline is an experimental gap, not circularity. The two self-citations (Valdenegro-Toro 2021; Valdenegro-Toro and Mori 2022) support only general UQ claims and are not load-bearing for the central accuracy claim. The implausible CoCoOp CUBirds value (New 8.04, F1 14.40 in Table 1) and the omission of CoCoOp from the main few-shot curves are correctness and benchmarking concerns, not a circular dependency, so the circularity score is 1.

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

The method introduces a new trainable cross-attention module but no new physical or conceptual entities. Its success depends on the frozen CLIP features, the capacity and initialization of the cross-attention layer, and the validity of MC dropout as an uncertainty approximation. Hyperparameters (dropout, learning rate, epochs, heads) are set by hand or borrowed from prior work without ablation.

free parameters (5)
  • Dropout rate = 0.2
    Chosen manually following Seoh's guidance on MC dropout; affects both regularization strength and uncertainty estimates.
  • Learning rate = 0.001
    Fixed SGD learning rate with cosine decay; not tuned per dataset.
  • Epoch schedule = 50/100/150 for 1/2-4/8-16 shots
    Heuristic adopted from CoOp and VPT; scales training time with number of examples.
  • Attention heads = 8
    Architecture choice for the cross-attention layer; no ablation.
  • MC samples = 100
    Number of dropout samples used for qualitative histograms (Figure 8); not systematically varied.
assumptions (4)
  • domain assumption CLIP's pretrained image and text encoders provide transferable features for fine-grained tasks.
    The method relies on frozen CLIP encoders as the foundation; if CLIP features are not discriminative for fine-grained classes, the cross-attention layer cannot compensate.
  • ad hoc to paper A single cross-attention layer can adapt text prompts to individual images without overfitting on few-shot data.
    The core architectural claim; only this layer is trainable, so its capacity must be sufficient but not excessive.
  • domain assumption Monte-Carlo Dropout provides a valid approximation of predictive uncertainty.
    Adopted from Gal and Ghahramani (2016); the paper relies on this to interpret entropy as confidence.
  • domain assumption Data augmentation (random resized crop and flip) preserves class identity for fine-grained recognition.
    Augmentation is applied during training; if it destroys subtle discriminative details, few-shot learning would suffer.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adaptive Prompt Tuning: Vision Guided Prompt Tuning with Cross-Attention for Fine-Grained Few-Shot Learning." pith.science (2026). https://pith.science/paper/MOAZCU7V

@misc{pith2026241214640,
  author       = {Pith},
  title        = {Pith review of: Adaptive Prompt Tuning: Vision Guided Prompt Tuning with Cross-Attention for Fine-Grained Few-Shot Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MOAZCU7V}},
  note         = {Machine review of arXiv:2412.14640}
}
read the original abstract

Few-shot, fine-grained classification in computer vision poses significant challenges due to the need to differentiate subtle class distinctions with limited data. This paper presents a novel method that enhances the Contrastive Language-Image Pre-Training (CLIP) model through adaptive prompt tuning, guided by real-time visual inputs. Unlike existing techniques such as Context Optimization (CoOp) and Visual Prompt Tuning (VPT), which are constrained by static prompts or visual token reliance, the proposed approach leverages a cross-attention mechanism to dynamically refine text prompts for the image at hand. This enables an image-specific alignment of textual features with image patches extracted from the Vision Transformer, making the model more effective for datasets with high intra-class variance and low inter-class differences. The method is evaluated on several datasets, including CUBirds, Oxford Flowers, and FGVC Aircraft, showing significant performance gains over static prompt tuning approaches. To ensure these performance gains translate into trustworthy predictions, we integrate Monte-Carlo Dropout in our approach to improve the reliability of the model predictions and uncertainty estimates. This integration provides valuable insights into the model's predictive confidence, helping to identify when predictions can be trusted and when additional verification is necessary. This dynamic approach offers a robust solution, advancing the state-of-the-art for few-shot fine-grained classification.

Figures

Figures reproduced from arXiv: 2412.14640 by the authors.

Figure 1
Figure 1. Overview of the proposed APT method. The method leverages CLIP’s image and text encoder (see Figure [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Diagram depicting the architecture of CLIP used [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Reliability plots help with qualitative evaluation [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Results of the few-shot learning set up. Our approach (red) is compared to the baseline CLIP results (purple), CoOp [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Expected Calibration Error (ECE) across number [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Calibration plots for different datasets showing the model’s performance in predicting correct class probabilities. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Confidence vs. Uncertainty plots for the Oxford Flowers, FGVC Aircraft, and CUBirds datasets show an inverse [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: The histograms on the left illustrate the predictions made over 100 samples with MC-dropout. The column on the [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: t-SNE plots of baseline prompts (red) versus fine-grained prompts (blue). The FGVC Aircraft dataset shows a larger [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 20 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence 'output.state := if if FUNCTION not #0 #1 if FUNCTION and 'skip pop #0 if FUNCTIO...

  2. [2]

    Z., Taha, T

    Alom, M. Z., Taha, T. M., Yakopcic, C., Westberg, S., Sidike, P., Nasrin, M. S., Hasan, M., Van Essen, B. C., Awwal, A. A., and Asari, V. K. (2019). A state-of-the-art survey on deep learning theory and architectures. electronics , 8(3):292

  3. [3]

    Blundell, C., Cornebise, J., Kavukcuoglu, K., and Wierstra, D. (2015). Weight uncertainty in neural network. In International conference on machine learning , pages 1613--1622. PMLR

  4. [4]

    Chen, T., Kornblith, S., Norouzi, M., and Hinton, G. (2020). A simple framework for contrastive learning of visual representations. In International conference on machine learning , pages 1597--1607. PMLR

  5. [5]

    Chen, Y., Liu, Z., Xu, H., Darrell, T., and Wang, X. (2021). Meta-baseline: Exploring simple meta-learning for few-shot learning. In Proceedings of the IEEE/CVF international conference on computer vision , pages 9062--9071

  6. [6]

    Z., Shu, G., Ortiz, E., et al

    Dehghan, A., Masood, S. Z., Shu, G., Ortiz, E., et al. (2017). View independent vehicle make, model and color recognition using convolutional neural network. arXiv preprint arXiv:1702.01721

  7. [7]

    Ding, K., Wang, J., Li, J., Shu, K., Liu, C., and Liu, H. (2020). Graph prototypical networks for few-shot learning on attributed networks. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management , pages 295--304

  8. [8]

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

Show all 33 references
  1. [9]

    and Ghahramani, Z

    Gal, Y. and Ghahramani, Z. (2016). Dropout as a bayesian approximation: Representing model uncertainty in deep learning. In international conference on machine learning , pages 1050--1059. PMLR

  2. [10]

    A., Hu, M., Malik, A

    Ganaie, M. A., Hu, M., Malik, A. K., Tanveer, M., and Suganthan, P. N. (2022). Ensemble deep learning: A review. Engineering Applications of Artificial Intelligence , 115:105151

  3. [11]

    and Fookes, C

    Goan, E. and Fookes, C. (2020). Bayesian neural networks: An introduction and survey. Case Studies in Applied Bayesian Data Science: CIRM Jean-Morlet Chair, Fall 2018 , pages 45--87

  4. [12]

    H \'e naff, O., Srinivas, A., Fauw, J., Razavi, A., Doersch, C., Eslami, S., and van den Oord, A. (2020). Data-efficient image recognition with contrastive predictive coding. arxiv. arXiv preprint arXiv:1905.09272

  5. [13]

    K., and Saxena, S

    Jena, B., Nayak, G. K., and Saxena, S. (2022). High-performance computing and its requirements in deep learning. In High-Performance Medical Image Processing , pages 255--288. Apple Academic Press

  6. [14]

    Jia, M., Tang, L., Chen, B.-C., Cardie, C., Belongie, S., Hariharan, B., and Lim, S.-N. (2022). Visual prompt tuning. In European Conference on Computer Vision , pages 709--727. Springer

  7. [15]

    and Yun, S.-Y

    Kim, S. and Yun, S.-Y. (2022). Calibration of few-shot classification tasks: Mitigating misconfidence from distribution mismatch. IEEE Access , 10:53894--53908

  8. [16]

    Lemley, J., Bazrafkan, S., and Corcoran, P. (2017). Deep learning for consumer devices and services: pushing the limits for machine learning, artificial intelligence, and computer vision. IEEE Consumer Electronics Magazine , 6(2):48--56

  9. [17]

    Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Doll \'a r, P., and Zitnick, C. L. (2014). Microsoft coco: Common objects in context. In Computer Vision--ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part V...

  10. [18]

    Maji, S., Kannala, J., Rahtu, E., Blaschko, M., and Vedaldi, A. (2013). Fine-grained visual classification of aircraft. Technical report

  11. [19]

    Miao, Y., Lei, Y., Zhou, F., and Deng, Z. (2024). Bayesian exploration of pre-trained models for low-shot image classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 23849--23859

  12. [20]

    Nakkiran, P., Kaplun, G., Bansal, Y., Yang, T., Barak, B., and Sutskever, I. (2021). Deep double descent: Where bigger models and more data hurt. Journal of Statistical Mechanics: Theory and Experiment , 2021(12):124003

  13. [21]

    Nguyen, A., Yosinski, J., and Clune, J. (2015). Deep neural networks are easily fooled: High confidence predictions for unrecognizable images. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

  14. [22]

    and Zisserman, A

    Nilsback, M.-E. and Zisserman, A. (2008). Automated flower classification over a large number of classes. In 2008 Sixth Indian conference on computer vision, graphics & image processing , pages 722--729. IEEE

  15. [23]

    Peng, Y., He, X., and Zhao, J. (2017). Object-part attention model for fine-grained image classification. IEEE Transactions on Image Processing , 27(3):1487--1500

  16. [24]

    W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al

    Radford, A., Kim, J. W., Hallacy, C., Ramesh, A., Goh, G., Agarwal, S., Sastry, G., Askell, A., Mishkin, P., Clark, J., et al. (2021). Learning transferable visual models from natural language supervision. In International conference on machine learning , pages 8748--8763. PMLR

  17. [25]

    Seoh, R. (2020). Qualitative analysis of monte carlo dropout. arXiv preprint arXiv:2007.01720

  18. [26]

    Snell, J., Swersky, K., and Zemel, R. (2017). Prototypical networks for few-shot learning. Advances in neural information processing systems , 30

  19. [27]

    Valdenegro-Toro, M. (2021). I find your lack of uncertainty in computer vision disturbing. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 1263--1272

  20. [28]

    and Mori, D

    Valdenegro-Toro, M. and Mori, D. S. (2022). A deeper look into aleatoric and epistemic uncertainty disentanglement. In 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) , pages 1508--1516. IEEE

  21. [29]

    N., Kaiser, ., and Polosukhin, I

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

  22. [30]

    Wah, C., Branson, S., Welinder, P., Perona, P., and Belongie, S. (2011). Technical Report CNS-TR-2011-001, California Institute of Technology

  23. [31]

    T., and Ni, L

    Wang, Y., Yao, Q., Kwok, J. T., and Ni, L. M. (2020). Generalizing from a few examples: A survey on few-shot learning. ACM computing surveys (csur) , 53(3):1--34

  24. [32]

    C., and Liu, Z

    Zhou, K., Yang, J., Loy, C. C., and Liu, Z. (2022a). Conditional prompt learning for vision-language models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 16816--16825

  25. [33]

    C., and Liu, Z

    Zhou, K., Yang, J., Loy, C. C., and Liu, Z. (2022b). Learning to prompt for vision-language models. International Journal of Computer Vision , 130(9):2337--2348

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.