SBERT UNO Sustainable Development Goals Model
| Property | Value |
|---|---|
| Author | Rodion |
| Model Type | Sentence Transformer |
| Vector Dimensions | 768 |
| Training Dataset Size | 20,000 records |
| Model URL | Hugging Face Hub |
What is sbert_uno_sustainable_development_goals?
This is a specialized SBERT (Sentence-BERT) model trained specifically on United Nations Sustainable Development Goals data. It transforms sentences and paragraphs into 768-dimensional dense vector representations, making it particularly effective for semantic similarity tasks related to sustainability and development goals.
Implementation Details
The model was trained on 16,000 records with 4,000 used for evaluation. It implements a sophisticated similarity calculation system based on class commonality, using three different cases for similarity measurement. The training utilized AdamW optimizer with a learning rate of 2e-05 and included warmup steps with linear scheduling.
- Utilizes MPNet architecture with mean pooling
- Implements CosineSimilarityLoss for training
- Supports maximum sequence length of 512 tokens
- Trained with batch size of 64
Core Capabilities
- Sentence and paragraph embedding generation
- Semantic similarity calculation
- Clustering of sustainability-related content
- Text classification for SDG-related tasks
Frequently Asked Questions
Q: What makes this model unique?
This model is specifically trained on UNO sustainable development goals data, making it particularly effective for sustainability-related NLP tasks. Its unique similarity calculation system considers class relationships in three different ways, providing more nuanced semantic understanding.
Q: What are the recommended use cases?
The model is ideal for tasks such as clustering sustainability-related documents, searching through SDG-related content, and measuring semantic similarity between sustainability initiatives. It's particularly useful for organizations working with UN Sustainable Development Goals.
