R

Roberta Base Use Qa Theseus Bg

Developed by rmihaylov
This is a multilingual Roberta model that can generate embeddings for Bulgarian sentences. Trained based on Sentence-BERT principles with Google's USE model as the teacher model.
Downloads 15
Release Time : 4/18/2022

Model Overview

This model is used to generate embeddings for Bulgarian sentences, suitable for tasks like sentence similarity calculation. Model compression was achieved through progressive module replacement techniques.

Model Features

Multilingual support
Supports generating sentence embeddings for both Bulgarian and English
Case sensitivity
The model is case-sensitive, treating words like 'bulgarian' and 'Bulgarian' as different vocabulary
Model compression
Model compression achieved through progressive module replacement techniques
Trained on translation pairs
Trained using Bulgarian-English parallel corpus, mapping translated sentences to the same vector space

Model Capabilities

Bulgarian sentence embedding generation
English sentence embedding generation
Sentence similarity calculation

Use Cases

Information retrieval
Q&A systems
Used to find the most relevant answers to questions
Examples demonstrate how to calculate similarity between questions and candidate answers
Text matching
Similar sentence identification
Identify semantically similar sentences
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase