Skip to main content

In-context Bayesian Learning Curve Extrapolation

Project description

Efficient Bayesian Learning Curve Extrapolation using Prior-Data Fitted Networks

This repository offers an implementation of LC-PFN, a method designed for efficient Bayesian learning curve extrapolation.

LC-PFN in action on Google colab and HuggingFace

Installation using pip:

pip install -U lcpfn

Usage

Try out the notebooks (require matplotlib) for training and inference examples.

NOTE: Our model supports only increasing curves with values in $[0,1]$. If needed, please consider normalizing your curves to meet these constraints. See an example in notebooks/curve_normalization.ipynb.

Reference

@inproceedings{
adriaensens2023lcpfn,
title={Efficient Bayesian Learning Curve Extrapolation using Prior-Data Fitted Networks},
author={Adriaensen, Steven and Rakotoarison, Herilalaina and Müller, Samuel and Hutter, Frank},
booktitle={Thirty-seventh Conference on Neural Information Processing Systems},
year={2023},
url={https://openreview.net/forum?id=xgTV6rmH6n}
}

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

lcpfn-0.1.3.tar.gz (29.2 kB view details)

Uploaded Source

Built Distribution

lcpfn-0.1.3-py3-none-any.whl (32.3 kB view details)

Uploaded Python 3

File details

Details for the file lcpfn-0.1.3.tar.gz.

File metadata

  • Download URL: lcpfn-0.1.3.tar.gz
  • Upload date:
  • Size: 29.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.9.16

File hashes

Hashes for lcpfn-0.1.3.tar.gz
Algorithm Hash digest
SHA256 4115cff9451229a7d10d7b56d8d7ad1a12a988f2f9066474581c43345a87588a
MD5 e50f1025e85b828c29ca37c59b0adf04
BLAKE2b-256 0e87d436494b97c38bd394fd7b8fa58651f853b0518aba92ecd823ab699b680c

See more details on using hashes here.

File details

Details for the file lcpfn-0.1.3-py3-none-any.whl.

File metadata

  • Download URL: lcpfn-0.1.3-py3-none-any.whl
  • Upload date:
  • Size: 32.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/5.1.1 CPython/3.9.16

File hashes

Hashes for lcpfn-0.1.3-py3-none-any.whl
Algorithm Hash digest
SHA256 7416cb91d189ffcd7ed42e5b5a551e3ee73896518c799e1ff062ed91e538c8a1
MD5 5071bf89f0b535743d02bddcced069d4
BLAKE2b-256 18f328d0a3b6c38d766638e1d0b2c0b1adb93c817646842a652a68174861edc5

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page