LaBSE-kbd-v0.1 / README.md
panagoa's picture
upd space emoji
df6bb03

A newer version of the Gradio SDK is available: 5.21.0

Upgrade
metadata
title: LaBSE Kbd V0.1
emoji: πŸ”
colorFrom: green
colorTo: blue
sdk: gradio
sdk_version: 5.20.0
app_file: app.py
pinned: false
short_description: LaBSE-kbd Cross-Lingual Sentence Similarity Demo

LaBSE-kbd Cross-Lingual Sentence Similarity Demo

This Gradio space demonstrates the capabilities of the panagoa/LaBSE-kbd-v0.1 model for cross-lingual sentence similarity. The model supports 109 languages including Kabardian, Russian, English, Turkish, and many others.

Features

  • Compare sentences between different languages (not limited to Russian and Kabardian)
  • Get similarity scores based on semantic meaning
  • Explore pre-defined examples
  • Visualize how well the model understands cross-lingual semantic similarity

How to Use

  1. Enter text in the first input box (any language, e.g., Russian, English, Turkish)
  2. Enter text in the second input box (any language, e.g., Kabardian, Russian, English)
  3. Click "Compute Similarity"
  4. View the similarity score

You can compare sentences between any of the 111 languages supported by the model. Try comparing the same meaning across different languages!

About the Model

LaBSE-kbd-v0.1 is a fine-tuned version of Google's Language-agnostic BERT Sentence Embedding (LaBSE) model, specifically enhanced to support the Kabardian language. It can map sentences from 111 languages (including Kabardian) to a shared vector space, enabling cross-lingual similarity comparisons.

Technical Details

  • Base Model: sentence-transformers/LaBSE
  • Fine-tuned for: Kabardian language
  • Architecture: BERT-based with sentence transformer approach
  • Embedding dimension: 768