Awesome Prior-Data Fitted Networks

A curated collection of papers, repositories, and resources on Prior-data Fitted Networks (PFNs).

What are PFNs?

As the name says, Prior-data Fitted Networks (PFNs) are a class of neural networks trained on synthetic datasets sampled from a prior distribution to directly approximate the posterior predictive distribution (PPD). They enable Bayesian prediction through in-context learning and have been applied across tabular data, time series, Bayesian optimization, symbolic regression, and beyond.

Example — TabPFN

A landmark success of the PFN framework is the Tabular Foundation Model (TabPFN) published in Nature: Accurate predictions on small data with a tabular foundation model. It showed that a TabPFN trained on over 100 million synthetic tabular tasks can outperform traditional ML models (like XGBoost, CatBoost, LightGBM) on a wide range of small real datasets.

• Key idea: Train a transformer on 100 million synthetic tabular tasks, enabling zero-training predictions on new datasets.
• Architecture: An adapted transformer encoder designed for two-dimensional tabular data, supporting categorical + numeric features, missing values, and heterogeneous distributions.
• Impact: TabPFN demonstrates that PFNs can act as foundation models for tabular data, achieving state-of-the-art accuracy on small-data benchmarks in seconds, outperforming conventional AutoML systems.

This work established PFNs as a new family of foundation models for structured data, analogous to LLMs for text.

Figure: Overview of TabPFN.

Highlights

Venue	Title	Code
ICLR 2023	TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Second	Code
Nature	Accurate predictions on small data with a tabular foundation model	Code

Foundations

Foundations and theoretical insights into PFNs, amortized inference, and Bayesian learning.

Venue	Title	Code
ICLR 2022	Transformers Can Do Bayesian Inference	Code
ICML 2023	Statistical Foundations of Prior-Data Fitted Networks	Code
NeurIPS 2023	Transformers as Statisticians: Provable In-Context Learning with In-Context Algorithm Selection	Code
ICML 2024	Is In-Context Learning in Large Language Models Bayesian? A Martingale Perspective	Code
ICML 2025	Can Transformers Learn Full Bayesian Inference in Context?	Code

Papers

Venue	Title	Code
arXiv 2026	PluRel: Synthetic Data unlocks Scaling Laws for Relational Foundation Models	Code
ICLR 2026	Foundation Models for Causal Inference via Prior-Data Fitted Networks	Code
ICLR 2026	GIT-BO: High-Dimensional Bayesian Optimization with Tabular Foundation Models	Code
ICLR 2026	Learning Posterior Predictive Distributions for Node Classification from Synthetic Graph Priors	Code
ICLR 2026	Using maximal information auxiliary variables to improve synthetic data generation based on TabPFN foundation models	—
20251106	TabPFN-2.5: Advancing the State of the Art in Tabular Foundation Models	Code
arXiv 2025	Orion-MSP: Multi-Scale Sparse Attention for Tabular In-Context Learning	Code
UnderReview @ ICLR 2026	PDE-PFN: Prior-Data Fitted Neural PDE Solver	—
UnderReview @ ICLR 2026	SR-PFN: Yet Another Sequential Recommendation Paradigm	—
UnderReview @ ICLR 2026	Time-Aware Prior Fitted Networks for Zero-Shot Forecasting with Exogenous Variables	—
UnderReview @ ICLR 2026	Transformers Can Do Bayesian Clustering	—
UnderReview @ ICLR 2026	DistPFN: Test-Time Posterior Adjustment for Tabular Foundation Models under Label Shift	—
UnderReview @ ICLR 2026	Task-Aligned Attention Retrieval for Scaling Tabular Foundation Models	—
UnderReview @ ICLR 2026	Large-Scale Pretraining Offers Modest Benefits for Tabular Transfer Learning	—
UnderReview @ ICLR 2026	MultiModalPFN: Extending Prior-Data Fitted Networks for Multimodal Tabular Learning	—
UnderReview @ ICLR 2026	RaBEL: Scale-Aware Radial-Basis Embeddings for Tabular Foundation Models	—
arXiv 2025	TabPFN: One Model to Rule Them All?	Code
arXiv 2025	TabImpute: Accurate and Fast Zero-Shot Missing-Data Imputation with a Pre-Trained Transformer	Code
arXiv 2025	Decoupled-Value Attention for Prior-Data Fitted Networks: GP Inference for Physical Equations	Code
arXiv 2025	Efficient Autoregressive Inference for Transformer Probabilistic Models	Code
arXiv 2025	GraphPFN: A Prior-Data Fitted Graph Foundation Model	Code
arXiv 2025	Turning Tabular Foundation Models into Graph Foundation Models	Code
arXiv 2025	Bringing Graphs to the Table: Zero-shot Node Classification via Tabular Foundation Models	Code
arXiv 2025	LimiX: Unleashing Structured-Data Modeling Capability for Generalist Intelligence	Code
arXiv 2025	Gradient Free Deep Reinforcement Learning With TabPFN	—
arXiv 2025	Clustering by Attention: Leveraging Prior Fitted Transformers for Data Partitioning	—
arXiv 2025	On Finetuning Tabular Foundation Models	Code
arXiv 2025	TabPFN-Wide: Continued Pre-Training for Extreme Feature Counts	Code
arXiv 2025	Chunked TabPFN: Exact Training-Free In-Context Learning for Long-Context Tabular Data	Code
arXiv 2025	From Tables to Time: How TabPFN-v2 Outperforms Specialized Time Series Forecasting Models	Code
arXiv 2025	Realistic Evaluation of TabPFN v2 in Open Environments	Code
SSRN 2025	MultiTabPFN: Codebook-based Extensions of TabPFN for High-Class-Count Tabular Classification	—
TCBBIO 2025	GPFN: Prior-Data Fitted Networks for Genomic Prediction	Code
RML @ NeurIPS 2025	Robust Multi-task Modeling for Bayesian Optimization via In-Context Learning	—
NeurIPS 2025	Mitra: Mixed Synthetic Priors for Enhancing Tabular Foundation Models	Code
NeurIPS 2025	Effortless, Simulation-Efficient Bayesian Inference using Tabular Foundation Models	Code
NeurIPS 2025	Do-PFN: In-context Learning for Causal Effect Estimation	Code
NeurIPS 2025	ConTextTab: A Semantics-Aware Tabular In-Context Learner	Code
NeurIPS 2025	ZEUS: Zero-shot Embeddings for Unsupervised Separation of Tabular Data	Code
NeurIPS 2025	TabDPT: An Open Tabular Foundation Model	Code
NeurIPS 2025	CausalPFN: Amortized Causal Effect Estimation via In-Context Learning	Code
NeurIPS 2025	A Closer Look at TabPFN v2: Understanding Its Strengths and Extending Its Capabilities	—
NeurIPS 2025	EquiTabPFN: A Target-Permutation Equivariant Prior Fitted Networks	Code
ECAI 2025	In-Context Decision Making for Optimizing Complex AutoML Pipelines	Code
FMSD @ ICML 2025	State-Space Models for Tabular Prior-Data Fitted Networks	Code
FMSD @ ICML 2025	Real-TabPFN: Improving Tabular Foundation Models via Continued Pre-training With Real-World Data	—
FMSD @ ICML 2025	From Tabular to Time Series: Can TabPFN Handle Mixed Data? A Study on PhysioNet	—
FMSD @ ICML 2025	Early Stopping Tabular In-Context Learning	—
FMSD @ ICML 2025	Explore the Time Series Forecasting Potential of TabPFN Leveraging the Intrinsic Periodicity of Data	Code
ICML 2025	Position: The Future of Bayesian Prediction Is Prior-Fitted	—
ICML 2025	Bayesian Neural Scaling Law Extrapolation with Prior-Data Fitted Networks	Code
ICML 2025	Zero-shot Meta-learning for Tabular Prediction Tasks with Adversarially Pre-trained Transformer	Code
ICML 2025	TabPFN-Unleashed: A Scalable and Effective Solution to Tabular Prediction	Code
ICML 2025	TabICL: A Tabular Foundation Model for In-Context Learning on Large Data	Code
ICML 2025	TabFlex: Scaling Tabular Learning to Millions with Linear Attention	Code
ICML 2025	FairPFN: A Tabular Foundation Model for Causal Fairness	Code
ICML 2025	Can Transformers Learn Full Bayesian Inference in Context?	Code
AISTATS 2025	Prior-Fitted Networks Scale to Larger Datasets When Treated as Weak Learners	Code
TMLR 2025	FoMo-0D: A Foundation Model for Zero-shot Tabular Outlier Detection	Code
AABI @ ICLR 2025	Uncertainty Quantification for Prior-Data Fitted Networks using Martingale Posteriors	—
FPI @ ICLR 2025	α-PFN: In-Context Learning Entropy Search	—
ICLR 2025	Mixture of In-Context Prompters for Tabular PFNs	—
ICLR 2025	KinPFN: Bayesian Approximation of RNA Folding Kinetics using Prior-Data Fitted Networks	Code
AAAI 2025	TimePFN: Effective Multivariate Time Series Forecasting with Synthetic Data	Code
ICLR 2025	MotherNet: Fast Training and Inference via Hyper-Network Transformers	Code
Nature 2025	Accurate predictions on small data with a tabular foundation model	Code
OpenReview	Attic: A New Architecture for Tabular In-Context Learning Transformers	Code
arXiv 2024	LaT-PFN: A Joint Embedding Predictive Architecture for In-context Time-series Forecasting	Code
arXiv 2024	Fine-tuned In-Context Learning Transformers are Excellent Tabular Data Classifiers	Code
arXiv 2024	Tokenize Features, Enhancing Tables: The FT-TabPFN Model for Tabular Classification	Code
xAI 2024	Interpretable Machine Learning for TabPFN	Code
TSALM @ NeurIPS 2024	Mamba4Cast: Efficient Zero-Shot Time Series Forecasting with State Space Models	Code
TRL @ NeurIPS 2024	GAMformer: Bridging Tabular Foundation Models and Interpretable Machine Learning	—
TRL @ NeurIPS 2024	Adapting TabPFN for Zero-Inflated Metagenomic Data	—
TRL @ NeurIPS 2024	Towards Localization via Data Embedding for TabPFN	—
TRL @ NeurIPS 2024	Exploration of autoregressive models for in-context learning on tabular data	—
NeurIPS 2024	TuneTables: Context Optimization for Scalable Prior-Data Fitted Networks	Code
NeurIPS 2024	Retrieval & Fine-Tuning for In-Context Tabular Models	Code
NeurIPS 2024	Drift-Resilient TabPFN: In-Context Learning Temporal Distribution Shifts on Tabular Data	Code
NeurIPS 2024	TabEBM: A Tabular Data Augmentation Method with Distinct Class-Specific Energy-Based Models	Code
ICL @ ICML 2024	TabMDA: Tabular Manifold Data Augmentation for Any Classifier using Transformers with In-context Subsetting	Code
ICML 2024	In-Context Freeze-Thaw Bayesian Optimization for Hyperparameter Optimization	Code
ME-FoMo @ ICLR 2024	In-Context Data Distillation with TabPFN	—
Blogposts @ ICLR 2024	What exactly has TabPFN learned to do?	Code
TRL @ NeurIPS 2023	Scaling TabPFN: Sketching and Feature Selection for Tabular Prior-Data Fitted Networks	Code
TRL @ NeurIPS 2023	Fine-Tuning the Retrieval Mechanism for Tabular Deep Learning	—
TRL @ NeurIPS 2023	TabPFGen -- Tabular Data Generation with TabPFN	Code
NeurIPS 2023	Efficient Bayesian Learning Curve Extrapolation using Prior-Data Fitted Networks	Code
NeurIPS 2023	ForecastPFN: Synthetically-Trained Zero-Shot Forecasting	Codef
ICML 2023	PFNs4BO: In-Context Learning for Bayesian Optimization	Code
ICML 2023	Statistical Foundations of Prior-Data Fitted Networks	Code
ICLR 2023	TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Second	Code
ICLR 2022	Transformers Can Do Bayesian Inference	Code

GitHub Repositories

Repository	Description
automl/PFNs	Canonical PFN implementation; synthetic task generation, Bayesian inference via transformers.
PriorLabs/TabPFN	Official Tabular PFN implementation (classification + regression)
PriorLabs/tabpfn-extensions	Extensions: interpretability, more classes, imputation, and analysis tools
PriorLabs/awesome-tabpfn	Community-curated list of TabPFN applications and papers
PriorLabs/tabpfn-time-series	Time-series adaptation of TabPFN
david-rundel/tabpfn_iml	Interpretability module for TabPFN (SHAP, feature attribution)
yandex-research/G2T-FM	Graph-to-Table Foundation Model: extend TabPFN to graph data
yandex-research/graphpfn	GraphPFN: PFN with graph priors and message-passing transformer
abacusai/ForecastPFN	PFN for zero-shot time-series forecasting
soda-inria/tabicl	a more scalable tabular foundation model
limix-ldm/LimiX	LimiX: a tabular foundation model generalizing TabPFN
autogluon/tabrepo	Living benchmark for tabular ML; includes foundation-model baselines
autogluon/autogluon	End-to-end AutoML framework supporting tabular, time-series, and multimodal data; often used as benchmark in PFN/TabPFN work

Other Applications

1. Predicting Early Outcomes of Prostatic Artery Embolization Using n-Butyl Cyanoacrylate Liquid Embolic Agent: A Machine Learning Study
2. Virtual Screening of Natural Anti-Senescent Compounds Based on Sq-TabPFN
3. From Rows to Yields: How Foundation Models for Tabular Data Simplify Crop Yield Prediction
4. Uncertainty-Aware Tabular Prediction: Evaluating VBLL-Enhanced TabPFN in Safety-Critical Medical Data
5. Early Fault Classification in Rotating Machinery With Limited Data Using TabPFN
6. Improved Ethereum Fraud Detection Mechanism with Explainable Tabular Transformer Model
7. Fast and Accurate Zero-Training Classification for Tabular Engineering Data
8. A Fast and Reliable Transformer-Based TabPFN Model for Liver Disease Diagnosis
9. Machine learning and radiomics for ventricular tachyarrhythmia prediction in hypertrophic cardiomyopathy: insights from an MRI-based analysis
10. Explainable Classification for Non-Small Cell Lung Cancer Based on Positron Emission Tomography Features and Clinical Data
11. Class-Imbalanced-Aware Adaptive Dataset Distillation for Scalable Pretrained Model on Credit Scoring
12. Fault Diagnosis of Slewing Bearing Using Audible Sound Signal Based on Time Generative Adversarial Network–TabPFN Method
13. A machine learning-based approach for individualized prediction of short-term outcomes after anterior cervical corpectomy
14. Foundation Models for Cybersecurity: A Comprehensive Multi-Modal Evaluation of TabPFN and TabICL for Tabular Intrusion Detection
15. TACO: TabPFN Augmented Causal Outcomes for Early Detection of Long COVID
16. Tabular Data with Class Imbalance: Predicting Electric Vehicle Crash Severity with Pretrained Transformers (TabPFN) and Mamba-Based Models
17. Tabular prior-data fitted network for urban air temperature inference and high temperature risk assessment
18. Kriging prior Regression: A Case for Kriging-Based Spatial Features with TabPFN in Soil Mapping
19. Uncertainty-Aware Tabular Prediction: Evaluating VBLL-Enhanced TabPFN in Safety-Critical Medical Data
20. Tabular foundation model for GEOAI benchmark problems BM/AirportSoilProperties/2/2025

Contributing

Contributions are welcome!

• Add papers, repos, or tutorials.
• Fix broken links.
• Suggest new categories.

License & Credits

This list aggregates publicly available PFN-related resources. Each work retains its own license and citation requirements. If you use these resources, please cite the corresponding paper and repository.