Show HN: Autofit2 – End-to-end pipeline for multilingual text classification
Category: ai-ml
Tags: few-shot-learning, text-classification, multilingual-nlp
Score: 6.0/10 (Innovation: 5, Technical: 6, Documentation: 7, Utility: 6)
Autofit2 is an automated pipeline for few-shot multilingual text classification that leverages setfit and SBERT embeddings to achieve high precision with minimal labeled data. Its interesting aspect is the combination of few-shot learning, multilingual support (50+ languages), and a reproducible pipeline with CO₂ tracking, making it accessible for low-resource language tasks.
Target audience: data scientists, NLP engineers, researchers
Repository: https://github.com/neospe/autofit2 · Python · MIT
View on Hacker News