TFM Library

A personal, curated knowledge base on tabular foundation models (TFMs) — the TabPFN family and everything around it. This repository keeps track of two things, in one place, for every research project that needs them:

1. All the interesting literature — every collected paper as PDF with a full-text extraction, plus two levels of written digest: per-paper summaries and a cross-paper synthesis of the whole field.
2. All the reference repositories — flat-text snapshots of the upstream implementations (TabPFN, its extensions, finetuning codebases, …) so that "how does the official code actually do X?" is always answerable with a text search, offline, against the exact code the papers shipped.

It is designed to be mounted as a git submodule inside multiple research projects (CreditPFN and future ones), so all of them share one consistently maintained copy instead of each drifting on its own. If you are an AI agent: read AGENTS.md before touching anything here.

Contents

Path	What you'll find
SYNTHESIS.md	Start here. The cross-paper synthesis of the TFM paradigm: lineage timeline (PFNs → TabPFN v1/v2/2.5/3 → scaling / adaptation / extensions), the design axes on which the field varies, recurring weaknesses, open frontiers, and a one-card-per-paper appendix.
SUMMARIES.md	The per-paper tour, chronological: venue, where each paper fits, what it contains, strengths/limitations, and per-project relevance notes.
papers/	The PDFs, named YYYY_Author_et_al._Title.pdf, with full text extractions under papers/text/ (same basename, .txt).
REPOSITORIES.md	What each code dump is, why it's kept, and what to grep for when building on it.
repositories/	The flat-text code snapshots themselves (made with gitingest).
CHANGELOG.md	Human-readable log of library updates — check it before bumping a project's submodule pin.
scripts/	Maintenance tools: refresh_repositories.py (re-snapshot all dumps), extract_paper_text.py (PDF → papers/text/).

How to browse

• "I want to understand the field" → read SYNTHESIS.md top to bottom.
• "What did paper X actually do?" → SUMMARIES.md, or grep the full text in papers/text/.
• "How does the official implementation handle Y?" → REPOSITORIES.md to pick the right dump, then grep repositories/*.txt by symbol name.

Using this library in a project

git submodule add <this-repo-url> tfm-library     # once, in the project root
git submodule update --init                        # after a fresh clone
git submodule update --remote tfm-library          # pull the latest library

Each project pins a specific commit of this library — reproducible by construction. After improving the library, bump the pin in each consumer.

Housekeeping notes

• repositories/TabPFN Wide.txt (~366 MB) exceeds GitHub's file limit and is gitignored; regenerate locally with python scripts/refresh_repositories.py.
• repositories/VSC Documentation.txt is KU Leuven VSC cluster documentation — not TFM literature, but shared here because every consuming project runs on that cluster.
• Keep this repository private: the PDFs are copyrighted publisher material, collected for private research use.

Consuming projects

Project	Mountpoint	Since
CreditPFN	tfm-library/	2026-07