ART-Group-it/KERMIT

🐸 KERMIT - A lightweight library to encode and interpret Universal Syntactic Embeddings

39
/ 100
Emerging

This tool helps natural language processing researchers and practitioners improve the performance of their Transformer models on linguistic tasks. It takes parsed sentence structures (parse trees) as input and provides an encoded representation that can be used to add explicit syntactic information to models like BERT. The output also allows for the interpretation and visualization of how much syntax influences the model's decisions, which is useful for analyzing model behavior.

No commits in the last 6 months.

Use this if you are working with natural language understanding and want to enhance your Transformer models by integrating explicit syntactic information from parse trees to achieve better performance and interpretability.

Not ideal if you are looking for a general-purpose natural language processing library that doesn't focus on detailed syntactic encoding or require explicit parse tree input.

natural-language-processing computational-linguistics deep-learning text-analysis model-interpretability
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 15 / 25

How are scores calculated?

Stars

58

Forks

9

Language

JavaScript

License

MIT

Last pushed

Jan 18, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/ART-Group-it/KERMIT"

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