Pith. sign in

REVIEW 5 major objections 6 minor 29 references

Merge-Friendly Post-Training Quantization for Multi-Target Domain Adaptation

T0 review · 5 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read The paper claims that quantization noise degrades model merging, and that HDRQ — a quantization scheme that flattens loss surfaces and anchors weights to the source — restores merging quality, gaining 4.21 mIoU at W4A4.

desk verdict HDRQ is a plausible, well-motivated engineering contribution with a useful empirical result, but its theoretical derivation is shaky and the ablation doesn't pin the gains to the two advertised regularizers. read the letter →

arxiv 2505.23651 v1 pith:QUJHXSLI submitted 2025-05-29 cs.LG cs.CV

Juncheol Shin , Minsang Seok , Seonggon Kim , Eunhyeok Park This is my paper
classification cs.LGcs.CV
keywords post-trainingquantizationmodelmergingmulti-targetdomainadaptationerrorbarrierHessianregularizationnoise-basedweightaveraginglosslandscapeflattening
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 identifies quantization noise as a hidden cause of poor model merging. When two domain-adapted models are each quantized before being averaged into one multi-target model, the discrete weight perturbations inflate the error barrier between them and the merged model loses accuracy. The authors propose HDRQ, a post-training quantization method that flattens the loss surface around each model by injecting quantization noise, pulls each quantized model back toward the shared source weights with an $\ell^2$ distance regularizer, and resolves rounding ambiguity during merging with noise sampling. In W4A4 semantic segmentation, merging models quantized with HDRQ outperforms merging models quantized with QDrop by 4.21 mIoU, while single-model accuracy stays comparable or better.

What carries the argument

The error barrier — the largest loss increase along the linear interpolation between two adapted models, relative to their mean loss — is the central object. HDRQ's theoretical story is that quantization adds uniform noise to weights, and a second-order Taylor expansion turns the barrier into terms controlled by the Hessians at the two endpoints and the merged point. Regularizing the Hessian (through injected noise), reducing the distance between the two models (through an $\ell^2$ pull toward the source), and resolving the rounding-parity ambiguity (through cosine-filtered noise sampling) each target one term in that expansion, with the Hessian's M-Lipschitz continuity providing the link between endpoint curvature and merged-point curvature.

What would settle it

Estimate the Hessian Lipschitz constant $M$ along the interpolation path between two domain-adapted models and check whether the gradient of one domain's loss at the other model is proportional to the interpolation vector; if the proportionality fails, the theoretical motivation for the distance regularizer collapses. A simpler check: merge two models after randomly permuting one model's weight indices to break the shared-basin assumption — if HDRQ still improves the merged result, its benefit is empirical rather than basin-based.

Watch

Extended reading notes

Core claim

The central claim is that quantization should be designed with a future merge in mind, not only with single-model accuracy in mind. Working in the setting of training-free multi-target domain adaptation, where separately fine-tuned models are fused by weight averaging, the paper shows that quantization noise raises the error barrier between the two adapted weight points, and that this rise can be countered at quantization time. HDRQ does this with three linked mechanisms: noise-based quantization whose expected loss includes a Hessian-penalizing second-order term, thereby flattening the loss landscape; a distance regularizer that keeps each quantized model close to the shared source weight, which by the triangle inequality keeps the two target models close to each other; and a noise-sampling rounding step, filtered by cosine similarity, that removes the ambiguity of rounding an odd midpoint when averaging two quantized weight grids. The experiments across semantic segmentation and Office-Home show the merged model's harmonic mean accuracy improves most at low bit-widths, which is exactly where quantization noise is largest.

Load-bearing premise

The load-bearing premise is that the loss landscape between the two adapted models is smooth enough that the Hessian changes slowly (an M-Lipschitz bound) and that the models truly occupy one shared basin, so one domain's gradient at the other model is a scaled copy of the interpolation direction; the paper does not measure either quantity, and if either fails the distance regularizer's theoretical justification loses its footing.

Editorial extensions

If this is right

  • Post-training quantization should be benchmarked on merging outcomes in addition to single-model accuracy, since the two metrics diverge sharply at low bit-widths.
  • The gap between HDRQ and QDrop grows from 1.19 to 4.21 mIoU as precision drops from W6A6 to W4A4, so merge-friendliness matters most where quantization noise is largest.
  • Noise-based quantization, used in prior single-model post-training quantization work, gains a new role: it can deliberately flatten the loss surface to make later weight averaging across models succeed.
  • The triangle-inequality argument means keeping quantized models close to their shared source is a sufficient proxy for keeping them close to each other, without ever seeing the other target domain.

Reading between the lines

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

  • Beyond the paper's setting, the same recipe — flatten the Hessian and anchor to a shared starting point — could plausibly improve merging for other post-hoc weight operations such as task arithmetic or model soups, though the paper only demonstrates it for domain-adapted models under simple weight averaging.
  • A testable extension the paper does not pursue is to measure the Lipschitz constant $M$ and the gradient-scaling scalar $k$ directly; small values would make the theory's bounds tight, while large values would leave the method's empirical success unexplained by the stated analysis.
  • The cosine-similarity-filtered sampling suggests a general principle for stochastic merges: when the merged weights are ambiguous, choose the sample whose merged direction best preserves the original interpolation geometry. This could be tried as a drop-in fix for any midpoint weight averaging, with or without quantization.
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 / 6 minor

Summary. The paper studies post-training quantization of separately domain-adapted models that are subsequently merged for multi-target domain adaptation. It proposes HDRQ, which combines noise-based Hessian regularization, a weight-distance regularization pulling each quantized model toward the source weights, and a noise-sampling rounding procedure designed to resolve rounding ambiguity during merging. The authors present a theoretical error-barrier analysis (Section 3) motivating these components and evaluate on GTA-to-Cityscapes/IDD semantic segmentation and Office-Home classification, reporting that HDRQ improves merged-model accuracy at low precision (e.g., +4.21 mIoU over QDrop at W4A4) while keeping single-model accuracy roughly comparable.

Significance. If substantiated, the paper identifies a practically important design axis for post-training quantization: current PTQ methods optimize per-model accuracy but ignore the compatibility of quantized models with subsequent weight merging. The empirical package is broad, using standard benchmarks and baselines, and the incremental ablation in Table 3 gives initial evidence that each proposed component contributes. However, the theoretical derivation as written is not rigorous, and the attribution of the headline merging gains to the two regularizers is not fully isolated from the advanced noise-sampling rounding in the main segmentation experiment. The work is a promising step rather than a settled result.

major comments (5)
  1. [Section 3.1, Eq. (4)] The displayed equality is not valid because the maximum of a sum is not generally equal to the sum of the maxima; only an inequality max[f+g] <= max[f] + max[g] holds without additional conditions. The subsequent minimization argument in Eq. (5) depends on this decomposition, so the theoretical analysis is not sound as written. Please rewrite the step as an inequality or prove the required exchange-of-maxima condition, and state how this affects the conclusions drawn from Eqs. (5)-(7).
  2. [Section 3.3, Eq. (7)] The scalar k is introduced as being proportional to the distance between θλ and θ2, but it is never defined, estimated, or bounded. The assertion that the Jacobian term ∇_θ L1(θ2) is a scaled copy of the interpolation direction is an unverified modeling assumption. Likewise, Eq. (6) invokes an M-Lipschitz Hessian without any attempt to estimate M or check the assumption on the actual adapted models. These assumptions are load-bearing because they justify both the distance regularization and the Hessian regularization. I would accept a clearly labeled heuristic motivation, but the current text presents these as derivation; please either measure/validate the assumptions (for example, gradient alignment or Hessian Lipschitz constants along the interpolation path) or explicitly downgrade the theoretical claims.
  3. [Sections 5.1-5.3, Tables 1-3] The attribution of the headline merging gain to the Hessian and distance regularizers is not fully established because the advanced noise-sampling rounding of Section 4.3 is not controlled for on the semantic segmentation task. The W4A4 +4.21 mIoU result is reported after merging 30 sampled weights, and the cosine-similarity-based selection is HDRQ-specific, but the text does not state whether the QDrop and BRECQ baselines also receive the same noise-sampling procedure. If the sampling is applied only to HDRQ, the comparison is unfair; if it is applied to all methods, that should be stated explicitly. In addition, Table 3 (Office-Home) is conducted in a setting where the paper says sampling is not applied, so it does not control for this component in the task where the headline gain is obtained.
  4. [Section 5.2, Table 2; Section 5.1, Table 1] The summary claims that HDRQ 'consistently outperforms' previous approaches and maintains 'comparable or superior' single-model performance, but the tables do not uniformly support this. At W8A8 in Table 2, the merged harmonic mean for HDRQ is below QDrop for all four source configurations (67.46 vs 67.91, 68.10 vs 68.75, 74.58 vs 74.79, 64.51 vs 64.86). In Table 1, the W4A4 single-model mIoU is lower for HDRQ for both targets (58.23 vs 58.92 on G->C and 48.68 vs 49.44 on G->I). The paper should report variance (standard deviations or multiple seeds) and either temper these broad claims or discuss the trade-off between single-model accuracy and merging quality.
  5. [All experimental tables] The experimental comparison is reported as point estimates without error bars, seed information, or standard deviations. This is especially problematic for the small gaps in Table 1 (for example, 61.54 vs 61.32 at W6A6) and for the semantic segmentation results that average over 30 sampled weights. Without variance estimates, the reader cannot assess whether the reported differences are statistically meaningful. Please add multiple independent runs or at least report the standard deviation of the merging metric across the sampled weights, and specify the number of independent quantization runs per condition.
minor comments (6)
  1. [Table 2, QDrop row for R->A,C,P] The W8A4 entry appears as '67.464.85', which looks like two numbers run together; please correct the formatting and re-verify the corresponding values.
  2. [Section 5.1] The text says the results are presented in 'Section 4.1', but the semantic segmentation results appear in Table 1 of Section 5.1; the cross-reference is wrong.
  3. [Eq. (11)] The merging formula for integer representations should explicitly state the assumed merged step size, such as (Δ1+Δ2)/2; otherwise the expression I_merged = floor((I1*Δ1 + I2*Δ2)/(Δ1+Δ2)) is not self-explanatory.
  4. [Figure 2] The axis labels v1 and v2 are not defined in the caption or text; please specify what these directions are (for example, principal components of the weight space).
  5. [Section 3.2] The sentence 'we minimize the second-order term in (7)' refers to Eq. (7), but Eq. (7) is introduced later in Section 3.3; please correct the cross-reference or renumber the equations.
  6. [General] No code or repository is mentioned; releasing the implementation would substantially improve reproducibility, especially given the many implementation details in the quantization pipeline.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the merging-compatibility claim is tested against external PTQ baselines, and the cited noise-quantization mechanism is independently supported and acknowledged as prior work.

full rationale

HDRQ's two regularizers are not predictions derived from fitted inputs. The Hessian regularization in Eq. (9) is the standard expected-loss expansion for additive quantization noise, explicitly attributed to prior work (Baskin et al. 2021; Défossez et al. 2022; Shin et al. 2023; Lin et al. 2023), including the authors' NIPQ, but the paper does not rely solely on that self-citation and the mechanism is external and independently published. The distance regularizer in Sec. 4.2 is a direct design objective motivated by the triangle inequality in Eq. (10); it is not dressed up as a measured prediction, and the downstream merging accuracy is measured on external benchmarks (GTA/Cityscapes/IDD and Office-Home) against BRECQ and QDrop. The 4.21 mIoU claim is an empirical comparison, not a quantity forced by a fitted constant. The ablation in Table 3 is incremental and omits cross-product controls (e.g., QDrop + distance only), so the attribution of gains to the Hessian versus distance term is not fully isolated, but that is an experimental-identification gap rather than circular reasoning. The unverified assumptions in Secs. 3.1 and 3.3 (M-Lipschitz Hessian bound in Eq. (6), single-basin gradient scaling) are correctness and rigor risks, not circular steps: even if those assumptions fail, the method's evaluation would stand or fall empirically. No self-citation is load-bearing for the core claim, and no equation is equivalent to its input by construction.

Assumptions & free parameters 2 free parameters · 7 assumptions · 0 invented entities

The central claim rests on a sequence of modeling assumptions about the loss landscape, the availability of source weights, and the validity of noise-based approximations. The main free parameter is the distance-regularization strength, fixed rather than tuned per task. No invented entities are introduced.

free parameters (2)
  • Distance regularization strength lambda = 5e-2 (fixed for all tasks)
    Controls the l2 penalty pulling adapted quantized weights toward the source weight; no sensitivity analysis is reported, and it directly influences the merging gains.
  • Number of noise samples for merging = 30 (semantic segmentation)
    Reported mean uses 30 sampled weights; the count is chosen by the authors and no sensitivity study is shown.
assumptions (7)
  • domain assumption Quantization error is well modeled as uniform additive noise with zero mean.
    Used to derive Eq. (3) and the Hessian-regularization effect in Eq. (9); prior works make the same approximation, but it ignores correlations and clipping.
  • domain assumption Loss Taylor expansion up to second order is accurate for quantization perturbations.
    Eq. (4) and Eq. (9) discard higher order terms; no verification is provided for 4-bit or 3-bit quantization where noise is large.
  • domain assumption The loss Hessian is M-Lipschitz continuous between the two adapted weights.
    Eq. (6) requires this to bound the Hessian at the merged point; the constant M is never estimated or checked.
  • domain assumption The two adapted models lie in a single basin that contains the source model.
    Section 3.3 relies on Li et al. (2024) that fine-tuned models from the same initialization share a basin; used to absorb first-order terms into the merged-point gradient.
  • domain assumption Gradients vanish at the converged weights and at the merged point.
    Section 3.1 drops all first-order terms because both weights have converged to the same loss; the merged point is not a local minimum in general.
  • domain assumption Batch statistics can be folded into weights so that merging needs only midpoint averaging.
    Section 5 states batch normalization layers are folded after quantization; this removes normalization statistic mismatch, a known source of merging error.
  • domain assumption Source weights are available at quantization time.
    Section 4.2 assumes access to the original source weights to compute distance regularization; this is not always true in source-free adaptation scenarios.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Merge-Friendly Post-Training Quantization for Multi-Target Domain Adaptation." pith.science (2026). https://pith.science/paper/QUJHXSLI

@misc{pith2026250523651,
  author       = {Pith},
  title        = {Pith review of: Merge-Friendly Post-Training Quantization for Multi-Target Domain Adaptation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QUJHXSLI}},
  note         = {Machine review of arXiv:2505.23651}
}
read the original abstract

Model merging has emerged as a powerful technique for combining task-specific weights, achieving superior performance in multi-target domain adaptation. However, when applied to practical scenarios, such as quantized models, new challenges arise. In practical scenarios, quantization is often applied to target-specific data, but this process restricts the domain of interest and introduces discretization effects, making model merging highly non-trivial. In this study, we analyze the impact of quantization on model merging through the lens of error barriers. Leveraging these insights, we propose a novel post-training quantization, HDRQ - Hessian and distant regularizing quantization - that is designed to consider model merging for multi-target domain adaptation. Our approach ensures that the quantization process incurs minimal deviation from the source pre-trained model while flattening the loss surface to facilitate smooth model merging. To our knowledge, this is the first study on this challenge, and extensive experiments confirm its effectiveness.

Figures

Figures reproduced from arXiv: 2505.23651 by the authors.

Figure 1
Figure 1. We propose a quantization scheme designed with future merging in mind. Our method ensures that networks are quantized to a more merge-friendly state, reducing the degradation induced by merging. target domain adaptation (Li et al., 2024) demonstrated that models fine-tuned for different target domains can be fused into a single general model via simple weight averaging, even in a training-free manner. This discovery… view at source ↗
Figure 2
Figure 2. Visualization of loss surfaces quantized with each method is shown. ResNet-50 adapted from Real domain to Clipart domain (R −→ C) is quantized to W4A8. HDRQ effectively regularize hessian with noise-based quantization, leading weights to flatter surface. wˆ = clamp(⌊ w ∆ ⌉, −2 b−1 , 2 b−1 − 1) · ∆, (8) where b denotes the bit width. We then sample the quanti￾zation noise ϵ from the quantization error w − wˆ and trai… view at source ↗
Figure 3
Figure 3. The distribution of harmonic mean accuracy for merging W4A8 quantized C−→ R and C−→ A models on the Office-Home dataset is presented. Our simple yet effective cosine similarity￾based method, denoted as Advanced, successfully filters out low￾quality weights, stabilizing merging outcomes. 4.2. Weight distance regularization Measuring the distance between separately adapted weights for each domain is challenging, as no… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

29 extracted references · 24 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Git re-basin: Merging models modulo permutation symmetries

    Ainsworth, S., Hayase, J., and Srinivasa, S. Git re-basin: Merging models modulo permutation symmetries. In The Eleventh International Conference on Learning Representations, 2023

    work page 2023

  3. [3]

    M., and Mendelson, A

    Baskin, C., Zheltonozhkii, E., Rozen, T., Liss, N., Chai, Y., Schwartz, E., Giryes, R., Bronstein, A. M., and Mendelson, A. Nice: Noise injection and clamping estimation for neural network quantization. Mathematics, 9 0 (17): 0 2144, 2021

    work page 2021

  4. [4]

    Bengio, Y., L \' e onard, N., and Courville, A. C. Estimating or propagating gradients through stochastic neurons for conditional computation. CoRR, abs/1308.3432, 2013. URL http://arxiv.org/abs/1308.3432

    arXiv 2013

  5. [5]

    The cityscapes dataset for semantic urban scene understanding

    Cordts, M., Omran, M., Ramos, S., Rehfeld, T., Enzweiler, M., Benenson, R., Franke, U., Roth, S., and Schiele, B. The cityscapes dataset for semantic urban scene understanding. In Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016

    work page 2016

  6. [6]

    Differentiable model compression via pseudo quantization noise

    D \'e fossez, A., Adi, Y., and Synnaeve, G. Differentiable model compression via pseudo quantization noise. Transactions on Machine Learning Research, 2022. ISSN 2835-8856

    work page 2022

  7. [7]

    K., McKinstry, J

    Esser, S. K., McKinstry, J. L., Bablani, D., Appuswamy, R., and Modha, D. S. Learned step size quantization. In International Conference on Learning Representations, 2020

    work page 2020

  8. [8]

    K., Roy, D., and Carbin, M

    Frankle, J., Dziugaite, G. K., Roy, D., and Carbin, M. Linear mode connectivity and the lottery ticket hypothesis. In III, H. D. and Singh, A. (eds.), Proceedings of the 37th International Conference on Machine Learning, volume 119 of Proceedings of Machine Learning Research, pp.\ 3259--3269. PMLR, 13--18 Jul 2020

    work page 2020

Show all 29 references
  1. [9]

    Unsupervised multi-target domain adaptation: An information theoretic approach

    Gholami, B., Sahu, P., Rudovic, O., Bousmalis, K., and Pavlovic, V. Unsupervised multi-target domain adaptation: An information theoretic approach. IEEE Transactions on Image Processing, 29: 0 3993--4002, 2020. doi:10.1109/TIP.2019.2963389

    2020

  2. [10]

    Deep residual learning for image recognition

    He, K., Zhang, X., Ren, S., and Sun, J. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2016

    2016

  3. [11]

    and Zheng, L

    Hou, Y. and Zheng, L. Source free domain adaptation with image translation, 2021

    2021

  4. [12]

    Hrda: Context-aware high-resolution domain-adaptive semantic segmentation

    Hoyer, L., Dai, D., and Van Gool, L. Hrda: Context-aware high-resolution domain-adaptive semantic segmentation. In European conference on computer vision, pp.\ 372--391. Springer, 2022

    2022

  5. [13]

    Training-free model merging for multi-target domain adaptation

    Li, W., Gao, H.-a., Gao, M., Tian, B., Zhi, R., and Zhao, H. Training-free model merging for multi-target domain adaptation. In European Conference on Computer Vision. Springer, 2024

    2024

  6. [14]

    \ BRECQ \ : Pushing the limit of post-training quantization by block reconstruction

    Li, Y., Gong, R., Tan, X., Yang, Y., Hu, P., Zhang, Q., Yu, F., Wang, W., and Gu, S. \ BRECQ \ : Pushing the limit of post-training quantization by block reconstruction. In International Conference on Learning Representations, 2021

    2021

  7. [15]

    Do we really need to access the source data? S ource hypothesis transfer for unsupervised domain adaptation

    Liang, J., Hu, D., and Feng, J. Do we really need to access the source data? S ource hypothesis transfer for unsupervised domain adaptation. In III, H. D. and Singh, A. (eds.), Proceedings of the 37th International Conference on Machine Learning, volume 119 of Proceedings of M...

    2020

  8. [16]

    Bit-shrinking: Limiting instantaneous sharpness for improving post-training quantization

    Lin, C., Peng, B., Li, Z., Tan, W., Ren, Y., Xiao, J., and Pu, S. Bit-shrinking: Limiting instantaneous sharpness for improving post-training quantization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.\ 16196--16205, June 2023

    2023

  9. [17]

    Source-free domain adaptation for semantic segmentation

    Liu, Y., Zhang, W., and Wang, J. Source-free domain adaptation for semantic segmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.\ 1215--1224, June 2021

    2021

  10. [18]

    Long, M., Zhu, H., Wang, J., and Jordan, M. I. Unsupervised domain adaptation with residual transfer networks. In Lee, D., Sugiyama, M., Luxburg, U., Guyon, I., and Garnett, R. (eds.), Advances in Neural Information Processing Systems, volume 29. Curran Associates, Inc., 2016

    2016

  11. [19]

    A., Van Baalen, M., Louizos, C., and Blankevoort, T

    Nagel, M., Amjad, R. A., Van Baalen, M., Louizos, C., and Blankevoort, T. Up or down? adaptive rounding for post-training quantization. In International Conference on Machine Learning, pp.\ 7197--7206. PMLR, 2020

    2020

  12. [20]

    T., Belal, A., Kiran, M., Dolz, J., Blais-Morin, L.-A., and Granger, E

    Nguyen-Meidine, L. T., Belal, A., Kiran, M., Dolz, J., Blais-Morin, L.-A., and Granger, E. Unsupervised multi-target domain adaptation through knowledge distillation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pp.\ 1339--1347, J...

    2021

  13. [21]

    R., Vineet, V., Roth, S., and Koltun, V

    Richter, S. R., Vineet, V., Roth, S., and Koltun, V. Playing for data: G round truth from computer games. In Leibe, B., Matas, J., Sebe, N., and Welling, M. (eds.), European Conference on Computer Vision (ECCV), volume 9906 of LNCS, pp.\ 102--118. Springer International Publis...

    2016

  14. [22]

    Nipq: Noise proxy-based integrated pseudo-quantization

    Shin, J., So, J., Park, S., Kang, S., Yoo, S., and Park, E. Nipq: Noise proxy-based integrated pseudo-quantization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.\ 3852--3861, June 2023

    2023

  15. [23]

    B., Ramesh, P., Hearn, T., and Hoffman, J

    Stoica, G., Bolya, D., Bjorner, J. B., Ramesh, P., Hearn, T., and Hoffman, J. Zipit! merging models from different tasks without training. In The Twelfth International Conference on Learning Representations, 2024. URL https://openreview.net/forum?id=LEYUkvdUhq

    2024

  16. [24]

    IDD: A Dataset for Exploring Problems of Autonomous Navigation in Unconstrained Environments

    Varma, G., Subramanian, A., Namboodiri, A., Chandraker, M., and Jawahar, C. IDD: A Dataset for Exploring Problems of Autonomous Navigation in Unconstrained Environments . In 2019 IEEE Winter Conference on Applications of Computer Vision (WACV), pp.\ 1743--1751, Los Alamitos, C...

    2019

  17. [25]

    Deep hashing network for unsupervised domain adaptation

    Venkateswara, H., Eusebio, J., Chakraborty, S., and Panchanathan, S. Deep hashing network for unsupervised domain adaptation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.\ 5018--5027, 2017

    2017

  18. [26]

    QD rop: Randomly dropping quantization for extremely low-bit post-training quantization

    Wei, X., Gong, R., Li, Y., Liu, X., and Yu, F. QD rop: Randomly dropping quantization for extremely low-bit post-training quantization. In International Conference on Learning Representations, 2022

    2022

  19. [27]

    Training-free pretrained model merging

    Xu, Z., Yuan, K., Wang, H., Wang, Y., Song, M., and Song, J. Training-free pretrained model merging. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.\ 5915--5925, June 2024

    2024

  20. [28]

    Multi-target unsupervised domain adaptation without exactly shared categories, 2018

    Yu, H., Hu, M., and Chen, S. Multi-target unsupervised domain adaptation without exactly shared categories, 2018

    2018

  21. [29]

    V., and Wang, J

    Zou, Y., Yu, Z., Kumar, B. V., and Wang, J. Unsupervised domain adaptation for semantic segmentation via class-balanced self-training. In Proceedings of the European Conference on Computer Vision (ECCV), September 2018

    2018

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.