Pith. sign in

REVIEW 3 major objections 6 minor 21 references

Real-TabPFN: Improving Tabular Foundation Models via Continued Pre-training With Real-World Data

T0 review · 3 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Continued pre-training of the TabPFNv2 checkpoint on 71 curated real-world tables raises mean normalized ROC-AUC from 0.954 to 0.976 across the 29 small classification datasets of the OpenML AutoML Benchmark.

desk verdict A credible, well-controlled continued pre-training result whose main claim hinges on a contamination audit the paper does not actually provide. read the letter →

arxiv 2507.03971 v1 pith:KYX3QDW7 submitted 2025-07-05 cs.LG cs.AIstat.MEstat.ML

Anurag Garg , Muhammad Ali , Noah Hollmann , Lennart Purucker , Samuel Müller , Frank Hutter This is my paper
classification cs.LGcs.AIstat.MEstat.ML
keywords tabularfoundationmodelscontinuedpre-trainingin-contextlearningTabPFNreal-worlddataL2-SPregularizationOpenMLAutoBenchmarkcontamination
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

This paper claims that a tabular foundation model pre-trained purely on synthetic data can be improved by a second, targeted pre-training phase on real-world tables. Starting from the TabPFNv2 checkpoint, the authors continue training for 20,000 steps on 71 curated OpenML and Kaggle datasets with a low learning rate and an L2-SP regularizer that keeps the weights near their starting point. On the 29 OpenML AutoML Benchmark classification tasks used in TabPFNv2's own evaluation, the resulting model, Real-TabPFN, raises mean normalized ROC-AUC from 0.954 to 0.976, a difference reported as significant at $p = 0.0045$ under a two-sided Wilcoxon signed-rank test. The authors also show that larger contexts during continued pre-training give larger gains and that a small curated corpus beats much larger web corpora such as CommonCrawl and GitTables. If correct, the result gives practitioners a stronger default model for small tabular classification at no extra inference cost, and it identifies curated real data, rather than more synthetic data, as the cheapest route to improving such models.

What carries the argument

The load-bearing mechanism is two-stage pre-training guarded by an L2-Starting-Point (L2-SP) regularizer. Stage 1 provides the TabPFNv2 transformer trained on more than 100 million synthetic tables; Stage 2 continues training on real tables while the penalty $\Omega(w) = \frac{\alpha}{2}\lVert w - w_0\rVert_2^2$, with $\alpha = 0.003$, penalizes squared distance from the initial weights $w_0$ and thereby curbs catastrophic forgetting. The low learning rate, linear warm-up with cosine annealing, batch size of one dataset, and the 60/40 context/query split let each real table enter as an in-context learning episode, so the improvement reflects a shift in the model's prior toward real-world table statistics rather than task-specific fine-tuning.

What would settle it

Publish the OpenML IDs and row/column hashes of the 71 training tables and check each of the 29 evaluation datasets, together with any superset, subsample, or renamed variant of it, against that list: if any evaluation task matches a training table, the 0.022 mean normalized ROC-AUC gain is contaminated and the claim falls. A second control is to retrain under the same recipe after deleting every training table from the same dataset family as any evaluation task and compare the resulting normalized ROC-AUC.

Watch

Extended reading notes

Core claim

The central discovery is that the synthetic-to-real gap in tabular foundation models can be closed by a short continued pre-training phase on a modest number of large, curated real-world tables. Keeping the TabPFNv2 architecture fixed, the authors continue pre-training on 71 datasets from OpenML and Kaggle, capped per table at 20,000 rows and 400,000 cells, with batch size 1, a 60% context / 40% query split, a learning rate of $3\times 10^{-7}$, and an L2-SP penalty with strength $\alpha = 0.003$ that pulls the weights back toward the synthetic checkpoint. Evaluated on the 29 AutoML Benchmark classification tasks below 10,000 samples and 500 features, Real-TabPFN improves mean normalized ROC-AUC from 0.954 to 0.976, and the paper reports a two-sided Wilcoxon signed-rank $p$ of 0.0045 against default TabPFNv2. The paper further reports that the gain grows with continued-pretraining context size, that OpenML and Kaggle sources are complementary (+0.019 and +0.015 alone, +0.022 combined), and that CommonCrawl and GitTables corpora yield smaller or negative gains, supporting curation over breadth.

Load-bearing premise

The claim stands on the assumption that the five-tier contamination filter fully removed every overlap between the 71 pre-training tables and the 29 evaluation tasks, even though both corpora come from OpenML and the paper reports no exclusion counts or per-dataset verification.

Editorial extensions

If this is right

  • Real-TabPFN becomes a stronger off-the-shelf baseline for small tabular classification than default TabPFNv2, with essentially the same inference cost (about 2.9 seconds per dataset in the paper's runs).
  • Continued-pretraining gains grow with context size, so the recipe has headroom: any increase in the 20,000-row or 400,000-cell cap should translate into further accuracy gains.
  • A small curated corpus of 71 large tables outperforms roughly 97,000 filtered CommonCrawl and 658 GitTables datasets, so corpus quality and table size dominate corpus breadth for this stage.
  • OpenML and Kaggle data are complementary: combining them gives +0.022 normalized ROC-AUC versus +0.015 or +0.019 alone, so pooling heterogeneous real sources is the better recipe.
  • Because Real-TabPFN beats every tuned baseline (AutoGluon, XGBoost, CatBoost, LightGBM, Random Forest, SVM, MLP, logistic regression) on mean normalized ROC-AUC, the default model becomes the recommended first call on small tabular problems.

Reading between the lines

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

  • The reported gain is an upper bound until leakage is ruled out: the paper asserts a five-tier contamination filter without reporting exclusion counts or per-dataset verification, and both the 71 training tables and the 29 evaluation tasks come from OpenML, so a released list of training dataset IDs and hashes would settle whether any evaluation task, superset, subsample, or renamed variant appears
  • The recipe is portable to future synthetic-pretrained tabular foundation models: the same low-rate, L2-SP-protected second phase could be applied to models that handle larger tables, and the paper's context-size curve predicts the gain would grow there.
  • The negative CommonCrawl result suggests a floor effect: small, noisy tables can actively degrade the learned prior, implying that table-size filtering should precede quality filtering when building tabular pre-training corpora, and that per-source ablation curves are a cheap screening test for candidate corpora.
  • A natural testable extension is regression and survival targets: the paper evaluates classification only, and it is an open question whether the 0.022 gain carries over when the loss and metric change.
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

3 major / 6 minor

Summary. The paper describes Real-TabPFN, obtained by continued pre-training of TabPFNv2 on 71 curated real-world datasets from OpenML and Kaggle, using a low learning rate, AdamW with cosine scheduling, and L2-SP regularization. Evaluation on 29 OpenML AutoML Benchmark classification tasks using 10-fold cross-validation reports mean normalized ROC-AUC improving from 0.954 (TabPFNv2 default) to 0.976, with a two-sided Wilcoxon signed-rank p = 0.0045. Additional ablations examine context size and data-source choices, claiming larger contexts and the union of OpenML and Kaggle corpora give the best results.

Significance. The paper's central empirical result, if contamination-free, is practically valuable: it demonstrates that a small, curated real-world corpus can improve a state-of-the-art synthetic-only tabular foundation model, yielding a stronger off-the-shelf baseline. The headline comparison is well controlled in design (same architecture, same base checkpoint, same evaluation protocol, and the authors commit to releasing the model weights), and the use of a Wilcoxon test on paired data is appropriate in principle. However, the main claim rests on an asserted but undocumented absence of train/evaluation overlap, and the secondary ablations lack any uncertainty quantification. The contribution is promising but not yet verified to the standard required for the claims made.

major comments (3)
  1. [Section 3, Appendices A-B] The five-tier contamination filter is described but its outcomes are never reported: there are no exclusion counts, no OpenML IDs for the 71 training datasets, and no per-dataset hash or name-matching results. Because both the pre-training and evaluation corpora are drawn from OpenML, and AMLB datasets are known to have supersets, subsamples, and duplicated uploads on OpenML, criterion (1) (only datasets with more than 10,000 samples) does not by itself rule out overlap. The statement 'We exclude any dataset from the pre-training data that does not meet these criteria' is an assertion, not evidence. This is load-bearing: a single leaked evaluation dataset, or a superset of one, could drive the reported difference between 0.954 and 0.976. Please provide the full training dataset list with OpenML IDs, the exclusion log (counts per filter stage), the hash-verification records against the 29 evaluation tasks, and a leakage sensitivity analysis (e.g., training only on the Kaggle subset and showing a similar result).
  2. [Section 5, Figures 4-6] All ablation results are single-seed point estimates with no error bars, repeated runs, or significance tests. For example, Figure 4 reports +0.010 for both 2048 and 4096 context sizes, and Figure 5 reports +0.015 (Kaggle), +0.019 (OpenML), and +0.022 (Both); these differences are small relative to the noise expected from a single continued pre-training run of 20,000 steps with batch size 1. Without multiple seeds or variance estimates, the ranking of context sizes and the conclusion that the union of sources is 'best' are not established. This does not invalidate the headline comparison, but it is load-bearing for the paper's secondary conclusions about why the method works.
  3. [Section 5, Figure 1] The Wilcoxon signed-rank test reported in Figure 1 is a paired test, but the manuscript does not state whether Real-TabPFN and TabPFNv2 were evaluated on the exact same 10-fold splits. If the fold splits differ between the two models, the pairing assumption is violated and the p-value of 0.0045 could overstate significance. Please state explicitly whether the fold indices are identical for both models; if they are not, rerun the comparison on shared folds or report an unpaired test such as the Mann-Whitney U test.
minor comments (6)
  1. [Figure 2 and Section 3] The text and figures use inconsistent corpus names: 'GitTable' versus 'GitTables' and 'CommonCrawl' versus 'Common Crawl'.
  2. [Figure 3 caption] The caption contains a typo: 'Comparsion' should be 'Comparison'.
  3. [Section 4] The L2-SP penalty is defined as the squared L2 norm of (w - w0); please call it the 'squared L2 norm' rather than 'L2 norm' to match the equation.
  4. [Appendix A] Provide OpenML dataset IDs for all OpenML entries; the current name-only list includes ambiguous entries (e.g., 'Bank marketing data set UCI' and 'Bank Marketing') that make it impossible for readers to verify the contamination filter.
  5. [Section 6] The sentence 'outperforms every other state-of-the-art baseline on all evaluated datasets' overstates the evidence in Table C, which shows the highest mean normalized ROC but not per-dataset dominance over every baseline; please rephrase to 'on average'.
  6. [References] The reference for Vanschoren et al. contains a truncated URL/DOI ('264119.'); please correct it.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the core result is a controlled empirical comparison with an independent external baseline and benchmark.

full rationale

The paper's central claim is empirical rather than derivational: Real-TabPFN is obtained by continuing pre-training of the public TabPFNv2 checkpoint on a curated corpus, and its improvement over TabPFNv2 is measured on the standard OpenML AutoML Benchmark suite. The reported gain in mean normalized ROC-AUC (0.954 to 0.976, Wilcoxon p = 0.0045) is an observed outcome of a controlled comparison in which architecture, starting checkpoint, and evaluation protocol are held fixed; it is not computed from, nor definitionally equal to, any fitted constant or hand-chosen hyperparameter in the paper. The L2-SP regularization strength, learning rate, and training corpus are choices, but none of these are fit to the evaluation metric in a way that would force the reported improvement. The reuse of the TabPFNv2 evaluation suite and baseline numbers from Hollmann et al. (2025) is normal benchmark reuse by overlapping authors: TabPFNv2 is a public model with released weights and code, the AutoML Benchmark is an external, standard suite, and the paper additionally runs both TabPFNv2 and Real-TabPFN itself under the same 10-fold protocol. No equation in Section 4 reduces the final accuracy to the pre-training objective or to a self-citation. The data-contamination concern raised in Section 3 is a leakage and external-validity risk rather than a circularity: the paper asserts a five-tier filter but reports no exclusion counts, hashes, or per-dataset verification evidence, so whether any evaluation dataset appears in the training corpus is unverified. However, without positive evidence of overlap, this is a correctness assumption, not a demonstrated circular reduction. No load-bearing step in the paper is equivalent to its own inputs by construction; therefore the circularity score is 0.

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

Real-TabPFN is an empirical claim, not a derivation, so the ledger is dominated by training choices and data assumptions rather than mathematical axioms. The free parameters are hand-chosen hyperparameters that shape the continued pre-training regime (regularization strength, learning rate, step count, context split, memory cap); none of them is fit to the evaluation metric, but none is ablated either, so their contribution to the gain is unknown. The key domain assumptions are the soundness of the TabPFNv2 base checkpoint, the effectiveness of the contamination filter, the representativeness of the curated real-world corpus, and the validity of the AMLB normalized-ROC protocol. No new entities are postulated; Real-TabPFN is a trained artifact with public weights, not an invented mechanism.

free parameters (5)
  • L2-SP regularization strength alpha = 0.003
    Hand-chosen; controls how far continued pre-training may deviate from the TabPFNv2 starting weights and directly shapes the training objective (Section 4).
  • Learning rate = 3e-7
    Hand-chosen for continued pre-training; no learning-rate sweep is reported (Section 4).
  • Number of training steps = 20,000
    Hand-chosen compute budget; no convergence or step-count ablation is shown (Section 4).
  • Context/query split ratio = 60% context, 40% query
    Chosen for the forward pass during continued pre-training; determines how many in-context examples the model sees per batch (Section 4).
  • Per-dataset cell cap = 400,000 cells
    Memory-driven cap that changes the effective sample count per dataset during continued pre-training (Section 4).
assumptions (4)
  • domain assumption TabPFNv2's synthetic-only base checkpoint is a sound starting point for continued pre-training
    Stage 2 inherits the base model's quality and prior; the paper does not re-validate the base model and relies on the authors' earlier Nature result (Sections 1 and 4).
  • domain assumption The five-tier contamination filter fully removes overlap between the 71 pre-training datasets and the 29 evaluation datasets
    Leakage-freedom rests on ID, name, shape, feature-name, and hash checks plus manual inspection; exclusion counts are not reported, so completeness is assumed (Section 3).
  • domain assumption The 29 AMLB evaluation datasets and the normalized ROC protocol faithfully measure downstream in-context learning quality
    Headline numbers use a suite and normalization chosen by the overlapping TabPFNv2 authors; raw ROC gains (0.929 to 0.932) are smaller than normalized gains (0.954 to 0.976) (Sections 3 and 5, Appendix C).
  • domain assumption The 71 curated OpenML and Kaggle datasets represent real-world tabular data
    Curation criteria beyond preprocessing are not specified; quality assessment is manual (Sections 3 and 4).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Real-TabPFN: Improving Tabular Foundation Models via Continued Pre-training With Real-World Data." pith.science (2026). https://pith.science/paper/KYX3QDW7

@misc{pith2026250703971,
  author       = {Pith},
  title        = {Pith review of: Real-TabPFN: Improving Tabular Foundation Models via Continued Pre-training With Real-World Data},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KYX3QDW7}},
  note         = {Machine review of arXiv:2507.03971}
}
read the original abstract

Foundation models for tabular data, like TabPFN, achieve strong performance on small datasets when pre-trained solely on synthetic data. We show that this performance can be significantly boosted by a targeted continued pre-training phase. Specifically, we demonstrate that leveraging a small, curated collection of large, real-world datasets for continued pre-training yields superior downstream predictive accuracy compared to using broader, potentially noisier corpora like CommonCrawl or GitTables. Our resulting model, Real-TabPFN, achieves substantial performance gains on 29 datasets from the OpenML AutoML Benchmark.

Figures

Figures reproduced from arXiv: 2507.03971 by the authors.

Figure 1
Figure 1. Per Dataset Normalized ROC Comparison of TabPFN (default) and Real-TabPFN (ours) on the 29 datasets from the OpenML AutoML Benchmark Datasets. Wilcoxon p refers to the two-sided Wilcoxon signed-rank test p value. tangible benefits, such as fewer hospital re-admissions or more precise credit-risk scoring, domains in which tabular data dominates. We enhance TabPFNv2’s in-context learning by continuing to pre-train it … view at source ↗
Figure 2
Figure 2. Distribution of dataset sizes (number of rows and fea￾tures) from various sources. The prevalence of smaller datasets in broad corpora like CommonCrawl and GitTable contrasts with the larger datasets from OpenML and Kaggle. Kaggle. We adopt the latter, manually curating 71 high￾quality datasets from OpenML and Kaggle [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. Mean Normalized ROC AUC Comparsion of Real￾TabPFN with all the default and the tuned versions of the baselines on the AutoMLBenchmark. Scores were normalized per dataset, with 1.0 representing the best and 0.0 the worst performance with respect to all baselines. each dataset at 400,000 total cells, adjusting the number of samples accordingly for datasets with too many attributes. 5. Experiments and Results We follow… view at source ↗
Figures from the paper (2 more)
Figure 6
Figure 6. Figure 6: Change in normalized ROC AUC as the training data source is varied. The changes are shown relative to the base TabPFNv2 model performance which was synthetically pre￾trained. gression (Cox, 1958) and Random Forest (Breiman, 2001) and subsequently removing noisy dataset…
Figure 5
Figure 5. Figure 5: Increase in normalized ROC AUC as the training data source is varied. The gains are shown relative to the base TabPFNv2 model performance which was synthetically pre￾trained. OpenML vs Kaggle. To understand the impact of our final curated 71 training data on model perf…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

21 extracted references · 3 canonical work pages

  1. [1]

    Web data commons - web table corpus 2015 / english-language relational subset,

    Bizer, C., Meusel, R., Lehmberg, O., Ritze, D., and Zope, S. Web data commons - web table corpus 2015 / english-language relational subset,

    work page 2015

  2. [8]

    org/abs/2207.12560

    URL https://arxiv. org/abs/2207.12560. Grinsztajn, L., Oyallon, E., and Varoquaux, G. Why do tree-based models still outperform deep learning on tab- ular data?,

    arXiv

  3. [11]

    Hollmann, N., M¨uller, S., Purucker, L., Krishnakumar, A., K¨orfer, M., Hoo, S

    URL https://arxiv.org/abs/2207.01848. Hollmann, N., M¨uller, S., Purucker, L., Krishnakumar, A., K¨orfer, M., Hoo, S. B., Schirrmeister, R. T., and Hutter, F. Accurate predictions on small data with a tabular foundation model. Nature, 637(8045):319–326,

    arXiv

  4. [15]

    doi: 10.1073/pnas.1611835114

    ISSN 1091-6490. doi: 10.1073/pnas.1611835114. URL http://dx.doi. org/10.1073/pnas.1611835114. Li, X., Grandvalet, Y ., and Davoine, F. Explicit in- ductive bias for transfer learning with convolutional networks,

    doi:10.1073/pnas.1611835114

  5. [17]

    URL https://arxiv.org/abs/ 2410.18164. Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V ., Thirion, B., Grisel, O., Blondel, M., Prettenhofer, P., Weiss, R., Dubourg, V ., Vanderplas, J., Passos, A., Cour- napeau, D., Brucher, M., Perrot, M., and Duchesnay, E. Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12:2825–2830,

    arXiv

  6. [18]

    Qu, J., Holzm¨uller, D., Varoquaux, G., and Morvan, M

    URL https://arxiv.org/abs/ 2406.17557. Qu, J., Holzm¨uller, D., Varoquaux, G., and Morvan, M. L. Tabicl: A tabular foundation model for in-context learn- ing on large data,

    arXiv

  7. [19]

    Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y ., Li, W., and Liu, P

    URL https://arxiv.org/ abs/2502.05564. Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y ., Li, W., and Liu, P. J. Exploring the limits of transfer learning with a unified text-to-text transformer. CoRR, abs/1910.10683,

    arXiv 1910

  8. [21]

    TabularFM: An Open Framework For Tabular Foundational Models

    URL https://arxiv. org/abs/2406.09837. Vanschoren, J., van Rijn, J. N., Bischl, B., and Torgo, L. Openml: networked science in machine learn- ing. SIGKDD Explorations, 15(2):49–60,

    work page Pith review arXiv

Show all 21 references
  1. [23]

    URL https://arxiv.org/abs/2005. 08314. 7 Real-TabPFN: Improving Tabular Foundation Models via Continued Pre-training With Real-World Data A. Training Datasets The following table lists the 71 datasets curated for continued pre-training, along with their source and access link....

    2005

  2. [2009]

    Dorogush, A

    doi: 10.1109/CVPR.2009.5206848. Dorogush, A. V ., Gulin, A., Gusev, G., Kazeev, N., Prokhorenkova, L. O., and V orobev, A. Fighting biases with dynamic boosting. CoRR, abs/1706.09516,

    2009

  3. [2013]

    URL http://doi.acm

    doi: 10.1145/2641190.2641198. URL http://doi.acm. org/10.1145/2641190.264119. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L. u., and Polosukhin, I. At- tention is all you need. In Guyon, I., Luxburg, U. V ., Bengio, S., Wallach, H., Fe...

  4. [2016]

    URL http:// arxiv.org/abs/1603.02754. Cox, D. R. The regression analysis of binary sequences. Journal of the Royal Statistical Society. Series B, 20(2): 215–242,

    arXiv

  5. [2017]

    D’souza, A., Swetha, M., and Sarawagi, S

    URL http://arxiv.org/abs/1706.09516. D’souza, A., Swetha, M., and Sarawagi, S. Synthetic tab- ular data generation for imbalanced classification: The surprising effectiveness of an overlap class. In Proceed- ings of the AAAI Conference on Artificial Intelligence , volume 39, p...

    arXiv

  6. [2018]

    Lin, T.-Y ., Maire, M., Belongie, S., Bourdev, L., Girshick, R., Hays, J., Perona, P., Ramanan, D., Zitnick, C

    URL https://arxiv.org/abs/ 1802.01483. Lin, T.-Y ., Maire, M., Belongie, S., Bourdev, L., Girshick, R., Hays, J., Perona, P., Ramanan, D., Zitnick, C. L., and Doll´ar, P. Microsoft coco: Common objects in con- text,

    arXiv

  7. [2019]

    URL http: //arxiv.org/abs/1910.10683. Tran, Q. M., Hoang, S. N., Nguyen, L. M., Phan, D., and Lam, H. T. Tabularfm: An open framework for tabular foundational models,

    1910 arXiv

  8. [2020]

    doi: 10.18653/v1/2020.acl-main.740

    Association for Computational Linguis- tics. doi: 10.18653/v1/2020.acl-main.740. URL https: //aclanthology.org/2020.acl-main.740/. Hollmann, N., M ¨uller, S., Eggensperger, K., and Hut- ter, F. Tabpfn: A transformer that solves small tab- ular classification problems in a second,

    2020 doi

  9. [2021]

    Jiang, M., Liu, K

    URL https://arxiv.org/ abs/2106.07258. Jiang, M., Liu, K. Z., Zhong, M., Schaeffer, R., Ouyang, S., Han, J., and Koyejo, S. Investigating data contamination for pre-training language models,

    arXiv

  10. [2022]

    URL https://arxiv.org/abs/ 2207.08815. 5 Real-TabPFN: Improving Tabular Foundation Models via Continued Pre-training With Real-World Data Gururangan, S., Marasovi ´c, A., Swayamdipta, S., Lo, K., Beltagy, I., Downey, D., and Smith, N. A. Don‘t stop pre- training: Adapt languag...

    arXiv

  11. [2023]

    Gardner, J., Perdomo, J

    URL https: //arxiv.org/abs/2310.07875. Gardner, J., Perdomo, J. C., and Schmidt, L. Large scale transfer learning for tabular data via language model- ing,

    arXiv

  12. [2024]

    Kelly, M., Longjohn, R., and Nottingham, K

    URL https: //arxiv.org/abs/2401.06059. Kelly, M., Longjohn, R., and Nottingham, K. The uci machine learning repository,

    arXiv

  13. [2025]

    Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L

    URL https://arxiv.org/abs/ 2405.13396. Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. Imagenet: A large-scale hierarchical image database. In 2009 IEEE Conference on Computer Vi- sion and Pattern Recognition, pp. 248–255,

    2009 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.