FFZwai/leyoai-cyber-ner

LeyoAI Cyber NER Model

English | 中文

English

Model Description

BERT-based NER model for cybersecurity entities (IP, Domain, CVE, Malware, etc.)

Installation

pip install transformers torch

Usage Scenarios

• Extract security entities
• Threat intelligence parsing
• Security log analysis

How to Use

from transformers import AutoTokenizer, AutoModelForTokenClassification

tokenizer = AutoTokenizer.from_pretrained("FFZwai/leyoai-cyber-ner")
model = AutoModelForTokenClassification.from_pretrained("FFZwai/leyoai-cyber-ner")

text = "CVE-2024-1234 affects example.com with malware TrickBot"
inputs = tokenizer(text, return_tensors="pt")
outputs = model(**inputs)

# Get predicted labels
predictions = outputs.logits.argmax(dim=-1)

Training Details

• Base Model: bert-base-uncased
• Training Data: 3,024 samples
• Validation Data: 648 samples
• Entity Types: IP, Domain, CVE, Malware, URL, etc.

---

中文

模型描述

基于 BERT 的网络安全实体识别模型（IP、域名、CVE、恶意软件等）

安装方法

pip install transformers torch

使用场景

• 提取安全实体
• 威胁情报解析
• 安全日志分析

使用方法

from transformers import AutoTokenizer, AutoModelForTokenClassification

tokenizer = AutoTokenizer.from_pretrained("FFZwai/leyoai-cyber-ner")
model = AutoModelForTokenClassification.from_pretrained("FFZwai/leyoai-cyber-ner")

text = "CVE-2024-1234 影响域名 example.com，涉及恶意软件 TrickBot"
inputs = tokenizer(text, return_tensors="pt")
outputs = model(**inputs)

# 获取预测标签
predictions = outputs.logits.argmax(dim=-1)

训练细节

• 基座模型: bert-base-uncased
• 训练数据: 3,024 samples
• 验证数据: 648 samples
• 实体类型: IP、域名、CVE、恶意软件、URL 等

---

About LeyoAI

LeyoAI is an AI MaaS platform by 杭州市上城区乐友信息服务工作室, providing specialized AI assistants for cybersecurity, video safety, workflow automation, and data analytics.

Website: https://leyoai.vercel.app