SciPhiIntroducing Triplex
State-of-the-art model for knowledge graph construction
Overview
Triplex is a finetuned version of Phi3-3.8B designed for creating knowledge graphs from unstructured data. It excels at extracting triplets - simple statements consisting of a subject, predicate, and object - from text or other data sources.
- 98% cost reduction for knowledge graph creation
- Outperforms GPT-4 at 1/60th the cost
- Enables local graph building with SciPhi's R2R
- Open-source and available on HuggingFace
Performance
Triplex significantly outperforms GPT-4 in accuracy for knowledge graph construction tasks while being much more cost-effective.
Resources
Usage
Python Example
Cost-Effective Solution
Triplex offers comparable results to GPT-4 at a fraction of the cost, making it an economical choice for knowledge graph construction.
Versatile Applications
Ideal for a wide range of tasks including text understanding, transformation, and knowledge graph construction.
Cost-Effective
Offers a 98% cost reduction for knowledge graph creation compared to traditional methods.
Advanced Features
Supports named entity recognition and relationship extraction for comprehensive knowledge graph construction.
Commercial Usage
The weights for the models are licensed under cc-by-nc-sa-4.0. However, we waive these restrictions for organizations with under $5M USD in gross revenue in the most recent 12-month period.
For organizations over this revenue limit or those wanting to remove GPL license requirements (dual-license), please contact our team for commercial licensing options.