R Observatory: Metrics for tabpfn version 0.4.0

Balamuta, James Joseph

doi:10.5281/zenodo.21843040

tabpfn

0.4.0Package record

Prior-Data Fitted Network Foundational Model for Tabular Data

0packages depend

4.1Kdownloads / year

53.0%test coverage

13/13flavors OK

Overview

About

Maintained by Max KuhnCRAN page ↗GitHub ↗Website ↗Issues ↗Declared funder: Posit Software, PBC

Provides a consistent API for classification and regression models based on the 'TabPFN' model of Hollmann et al. (2025), "Accurate predictions on small data with a tabular foundation model," Nature, 637(8045) doi:10.1038/s41586-024-08328-6. The calculations are served via 'Python' to train and predict the model.

Changes in 0.4.0

Documented how to pin a version of one of the Python tabpfn library's own dependencies, such as torch or numpy, when a new release of that dependency breaks a tabpfn version that worked before. Use extra_packages in install_tabpfn() for the persistent environment, or reticulate::py_require() for the ephemeral one. See the "Pinning dependencies" section in ?install_tabpfn (#40).
The training_set_limit argument of tab_pfn() now defaults to Inf, so all of your data is used. It previously sampled anything larger down to 10,000 rows without saying so.
tab_pfn() no longer checks your data against the model's limits; the Python library does, and its error names the model in use. Limits now follow the version you chose rather than one fixed set.
Fixed a failure that stopped models from fitting at all on some setups, reported as AttributeError: property 'stream' of '_StderrHandler' object has no setter (#40). The package no longer uses reticulate::py_capture_output(), which reassigns the streams of Python's logging handlers and errors when one of them is read-only.
Failures coming from the Python library are now reported as R errors. The message keeps what the library said and drops the wrapper around it, and tab_pfn() adds what to do about it in R, such as lowering training_set_limit. This covers predict() as well.
?tab_pfn now lists each model version's limits, including the lower one that applies on a CPU, and its examples use current versions rather than earlier ones.
Test snapshots now record TabPFN v3.5, the model the Python library currently defaults to. They will need updating again whenever that default moves.
CI pins torch to 2.13.0. The 2.14.0 release triggers the _StderrHandler failure above, and the pin stays until that is settled upstream.

From NEWS.md in the release

Install

R (CRAN)pak (dev)Ubuntu (r2u)Fedora (COPR)openSUSE (OBS)

install.packages("tabpfn", repos = "https://cloud.r-project.org/")

Installs the current CRAN release. Prebuilt binaries on Windows and macOS.

if (!requireNamespace("pak", quietly = TRUE))
  install.packages("pak", repos = "https://cloud.r-project.org/")
pak::pak("tidymodels/tabpfn")

Development version from tidymodels/tabpfn. Installs pak first if it is not already available.

Release

# enable cranapt (r2u) if it is not already set up, then install
grep -rqs r2u.stat.illinois.edu /etc/apt/sources.list /etc/apt/sources.list.d/ \
  || curl -fsSL https://raw.githubusercontent.com/eddelbuettel/r2u/master/inst/scripts/add_cranapt_noble.sh | sudo bash
sudo apt install -y r-cran-tabpfn

Fast Ubuntu binaries via r2u (no compilation). The cranapt setup runs once and is skipped when it is already enabled. We list this channel from recorded r2u download activity rather than from a live index of the repository, so treat it as last known rather than guaranteed.

bspm (used by r2u) also needs make: sudo apt install make

# enable the iucar/cran COPR if it is not already present, then install
ls /etc/yum.repos.d/*iucar*cran*.repo >/dev/null 2>&1 \
  || sudo dnf copr enable -y iucar/cran
sudo dnf install -y R-CRAN-tabpfn

RPMs from the iucar/cran COPR, rebuilt from CRAN.

For install.packages() integration in R, also: sudo dnf install R-CoprManager

# add the autoCRAN OBS repo (Tumbleweed) if it is missing, then install
zypper -q lr | grep -qi autoCRAN \
  || sudo zypper addrepo https://download.opensuse.org/repositories/devel:/languages:/R:/autoCRAN/openSUSE_Tumbleweed/devel:languages:R:autoCRAN.repo
sudo zypper install -y R-tabpfn

RPMs auto-built from CRAN on openSUSE OBS (Tumbleweed). We list this channel from recorded autoCRAN download activity rather than from a live index of the repository, so treat it as last known rather than guaranteed.

Health

CRAN checks

13OK

CRAN check page ↗

Slowest check: 13.8 min · r-release-windows-x86_64

Status changes

OK13 OK2026-08-04
NOTE12 OK, 1 NOTE2026-08-01
OK7 OK2026-03-19

Advisories

none known

How this is measured·How advisories are found

Documentation

In release 0.4.0

READMEYes · 472 wordsVignettesNoNEWSYes · 100% structured · mentions 0.4.0Citation fileNoWebsiteYes · pkgdown

Help pages: 10 (2 internal)includes a package overview page
Examples that run: 43%3 run in a plain check (1 of them under a condition) · 1 only when donttest examples are enabled (R CMD check --as-cran) · 3 never run
Exported functions with examples: 75%
Documented parameters: 100%
Return-value docs: 88%
Help pages citing references: 20%

How these are measured

Downloads

4.1K

CRAN downloads in the past year

Rank #5,390 · ~11/day · ~344/mo

Releases R versions

Daily download trend is not available in this view yet.

72030 days

1.5K90 days

4.1K1 year

Compare downloads with other packages →

Also on61 r2u1 autocran255 conda_forge

How downloads are counted

Repository

tidymodels/tabpfn ↗Organization: tidymodels

35Stars

5Forks

3Open issues

0Open PRs

4Releases

120Commits

5Commit identities

License file reads as Apache-2.0 · 120 commits · Last activity 2026-09-25 · +2.9% stars, 30d · Default branch at 0.4.0.9000, ahead of CRAN 0.4.0 (checked 2026-10-18)

Responsiveness

How these are sampled

PRs merged (all time): 94%
Median days to close, 50 most recently updated closed issues: 8
Median days to close, 50 most recently updated closed PRs: 1

Stars over time

2025-07-30 · 122026-10-13 · 35

How these are measured

PracticesBeta

9 development-tooling and community-health practices detected across 5 families in the upstream repository

Checks run against github.com/tidymodels/tabpfn on 2026-10-18.

Continuous integration (2)

GitHub ActionsR CMD check workflow

The R CMD check workflow runs on Linux, macOS and Windows, and on R-devel

Governance and community (1)

Code of conduct

Docs source (4)

README.Rmdpkgdown

GitHub Pages site last deployed 2026-09-25 · The repository has a site deploy workflow

pkgdown site last built 2026-09-25

Coverage (1)

Coverage reporting (Codecov)

Show all practicesHide 1 practice

Lint, format, editor (1)

RStudio project

How this is detected·Detection ruleset v3 (2026-09-26)

Dependencies

Declared dependencies

17 external dependencies (excludes base and recommended)

Installing tabpfn adds 26 packages (17 with compiled code, 11 MB of source tarballs). Heaviest: Rcpp 1.8 MB, reticulate 1.6 MB, jsonlite 1.0 MB.

Depends (1)

R >= 4.1.0

Imports (9)

cli dplyr generics hardhat jsonlite purrr reticulate rlang tibble

LinkingTo (0)

none

Suggests (9)

covr ggplot2MASSmodeldata recipes rstudioapi spelling testthat withr

Enhances (0)

none

Dependency license breakdown →

Reverse dependencies

0direct

0indirect

Nothing depends on this yet.

How these are counted

Code & Tests

People & History

People (3)

Max Kuhn

Author, Maintainer since 0.1.0 (2026)

Edgar Ruiz

Author · added in 0.4.0

Posit Software, PBC

Copyright holder, Funder

Releases (4)

Pick two releases to compare their code metrics.

0.4.0Latest
record
2026-09-25 · current release · diff ↗
0.3.0
record
2026-09-01 · diff ↗ ·
added 2 exports
install_tabpfn, tabpfn_initialize
R
R 4.6.0 released · 2026-04-24
0.2.0
record
2026-03-18 · diff ↗ ·
added 2 exports
tabpfn_download_models, tabpfn_list_versions
0.1.0
record
2026-03-18
R
R 4.5.0 released · 2025-04-11

Package metadata

First published: 2026-03-18
Total releases: 4 / 1 yrs
License: Apache License (>= 2)
Minimum R: ≥ 4.1.0
Download size: 220 KB

Cite

Cite this package

Run in R for the authors' preferred citation:

citation("tabpfn")

Format

Kuhn, M., Posit Software, PBC, & Ruiz, E. (2026). tabpfn: Prior-Data Fitted Network Foundational Model for Tabular Data (Version 0.4.0) [Computer software]. https://doi.org/10.32614/CRAN.package.tabpfn

This package has no citation file, so citation() returns this entry built from DESCRIPTION.

Cite the R Observatory

For a number measured here: a download total, a coverage figure, an archival date.

Format

APA

Balamuta, J. J. (2026). R Observatory: Metrics for tabpfn version 0.4.0 [Data set]. HJJB, LLC. Data release v2026-10-21. https://doi.org/10.5281/zenodo.21843040

From data release v2026-10-21, which the citation names so these numbers can be found later. More on citing and the projects behind them.

Report a problem with this page →