# High precision F1

This model is a fine-tuned LLM router classifier based on ModernBERT-large, used to distinguish between complex queries (using large language models) and simple queries (using small language models).

A Russian multi-class sentiment classification model fine-tuned based on RuBERT, supporting neutral, positive, and negative sentiment classifications.

This is a Russian short-text sentiment classification model fine-tuned based on RuBERT-tiny2, supporting neutral, positive, and negative sentiment analysis.

A large-scale Arabic named entity recognition (NER) model built on a completely new dataset, capable of recognizing 9 different types of entities

A spelling error detection model based on the DistilBERT architecture, used to identify spelling errors in text