nphdang/FS-BBT
Black-box Few-shot Knowledge Distillation
This project helps machine learning practitioners create smaller, more efficient AI models when they only have a few examples of data and can't directly access the internal workings of a larger, existing model. It takes a limited set of unlabeled images and a 'black-box' teacher model to produce a lightweight student model capable of classifying images. It's designed for data scientists, ML engineers, or researchers working with sensitive or proprietary large models.
No commits in the last 6 months.
Use this if you need to compress a large image classification model into a smaller one, but you only have a small dataset for training and the original model's details are confidential or inaccessible.
Not ideal if you have abundant labeled data and full access to the internal parameters of the large 'teacher' model you wish to distill.
Stars
14
Forks
1
Language
Python
License
—
Category
Last pushed
Jul 19, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/nphdang/FS-BBT"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
Guang000/Awesome-Dataset-Distillation
A curated list of awesome papers on dataset distillation and related applications.
dkozlov/awesome-knowledge-distillation
Awesome Knowledge Distillation
SforAiDl/KD_Lib
A Pytorch Knowledge Distillation library for benchmarking and extending works in the domains of...
SakurajimaMaiii/ProtoKD
[ICASSP 2023] Prototype Knowledge Distillation for Medical Segmentation with Missing Modality
HikariTJU/LD
Localization Distillation for Object Detection (CVPR 2022, TPAMI 2023)