dbolya/parc

A benchmark suite for Scalable Diverse Model Selection for Accessible Transfer Learning from our NeurIPS 2021 paper.

28
/ 100
Experimental

This tool helps machine learning practitioners evaluate how well different pre-trained models can be adapted to new, specific image recognition tasks. It takes a collection of diverse pre-trained image models and a set of "probe" datasets, then provides a score indicating each model's suitability for transfer learning. It's designed for researchers and engineers who need to select the most effective pre-trained models for their particular computer vision applications without extensive trial and error.

No commits in the last 6 months.

Use this if you need a systematic way to compare and select pre-trained image models for transfer learning to ensure they are well-suited for your specific downstream tasks.

Not ideal if you are looking for a simple, out-of-the-box solution to fine-tune a single pre-trained model for a new task without needing to compare multiple options.

transfer-learning model-selection computer-vision deep-learning image-recognition
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 8 / 25
Community 14 / 25

How are scores calculated?

Stars

15

Forks

3

Language

Python

License

Last pushed

Dec 14, 2022

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/dbolya/parc"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.