Ihor/OpenBioLLM-Text2Graph-8B

TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kTool Calling:SupportedPublished:Jan 3, 2025License:apache-2.0Architecture:Transformer Open Weights Cold

Ihor/OpenBioLLM-Text2Graph-8B is an 8 billion parameter biomedical annotation model developed by Ihor Stepanov, designed for generating named entity annotations from unlabeled biomedical text. This model specializes in high-throughput, cost-efficient synthetic biomedical Named Entity Recognition (NER) data generation. It functions as the synthetic annotation backbone for GLiNER-BioMed models, enabling the identification of entities and inference of directed relationships in biomedical text.

Loading preview...

OpenBioLLM-Text2Graph-8B: Biomedical Text-to-Graph Annotation

Ihor/OpenBioLLM-Text2Graph-8B is an 8 billion parameter model specifically engineered for advanced biomedical text processing. Introduced in the paper "GLiNER-BioMed: A Suite of Efficient Models for Open Biomedical Named Entity Recognition", its core function is to transform raw biomedical text into structured graph data by performing Named Entity Recognition (NER) and Relation Extraction (RE).

Key Capabilities

  • Biomedical NER & RE: Identifies unique, contextually relevant entities and infers directed relationships between them from biomedical text.
  • JSON Output: Structures annotated data in a clear JSON format, including entity IDs, text, types, and relationships (head, tail, type).
  • High-Throughput Data Generation: Designed for efficient, large-scale synthetic biomedical NER data generation, serving as a crucial component for GLiNER-BioMed models.
  • Contextual Understanding: Ensures that all extracted relations exist only between annotated entities and are human-readable, reflecting real-world biomedical concepts.

Good For

  • Synthetic Data Creation: Generating large volumes of annotated biomedical data for training and evaluating other NER/RE models.
  • Biomedical Information Extraction: Automatically extracting structured information from scientific papers, clinical notes, and other biomedical texts.
  • Knowledge Graph Construction: Building or enriching biomedical knowledge graphs by identifying entities and their interconnections.
  • Research & Development: Supporting research in bioinformatics and natural language processing by providing a robust tool for biomedical text annotation.

Finetunes
1 models