README.md

metadata

pipeline_tag: text2text-generation
tags:
  - STIX
  - NER
  - RE
  - CTI
  - cyber threat intelligence
metrics:
  - f1: 0.34163032380169717
  - recall: 0.37442129629629634
  - precision: 0.3141203703703703

Model Card for Model ID

Model Details

Model Description

• Model to extract Relations from cyber threat intelligence(CTI) text.

• Standalone Model:

• Developed by: Leon Lukas

• Model type: seq2seq

• **Language(s) (NLP): English

• **Finetuned from model : mrmoor/cti-t5-RE-NYT (T5 model trained on NYT RE)

Metrics test set

• precision: 0.3141203703703703
• recall: 0.37442129629629634,
• f1 : 0.34163032380169717

Model Sources [optional]

• Repository: https://github.com/l0renor/Relation-Extraction-and-Knowledge-Graph-Generation-on-MISP-Event-Reports
• Paper [optional]: https://github.com/l0renor/Relation-Extraction-and-Knowledge-Graph-Generation-on-MISP-Event-Reports