# Unsupervised training

DINOv2 is a visual feature extraction model developed by Facebook Research that generates robust visual features without supervised learning.

The Pearl Small Model is a lightweight string embedding model specifically designed for string semantic similarity calculation, generating high-quality embedding vectors for tasks such as string matching and entity retrieval.

CompCSE and SimCSE are contrastive learning-based sentence embedding models for calculating sentence similarity.

This is a Korean small RoBERTa model based on sentence-transformers, capable of mapping sentences and paragraphs into a 256-dimensional dense vector space, suitable for tasks such as clustering or semantic search.

An unsupervised sentence encoder based on RoBERTa, utilizing DropHead technology to enhance feature space, suitable for sentence similarity calculation.

This is a semantic vector model based on the sentence-transformers framework, capable of mapping Dhivehi sentences and paragraphs into a 768-dimensional dense vector space, suitable for tasks such as text clustering or semantic search.