lancopku/DynamicKD

Code for EMNLP 2021 main conference paper "Dynamic Knowledge Distillation for Pre-trained Language Models"

36
/ 100
Emerging

When developing specialized AI models for language tasks, you often start by training a large, powerful "teacher" model and then distilling its knowledge into a smaller, faster "student" model. This project provides methods to dynamically adjust how that knowledge transfer happens. It takes your pre-trained teacher and student language models and helps the student learn more effectively, resulting in a more efficient, high-performing smaller model.

No commits in the last 6 months.

Use this if you are a machine learning engineer or researcher looking to create smaller, more efficient natural language processing models without sacrificing too much performance compared to large pre-trained models.

Not ideal if you are looking for a ready-to-use, off-the-shelf NLP model without needing to engage in model training or fine-tuning.

natural-language-processing model-optimization deep-learning-research language-model-compression
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 13 / 25

How are scores calculated?

Stars

41

Forks

6

Language

Python

License

MIT

Last pushed

Aug 09, 2022

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/nlp/lancopku/DynamicKD"

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