# zjunlp/DeepKE

[EMNLP 2022] An Open Toolkit for Knowledge Graph Extraction and Construction

Repository: https://github.com/zjunlp/DeepKE
Canonical: https://ross.abutalabs.com/products/deepke
Homepage: http://deepke.zjukg.cn/
Language: Python
License: MIT
License Family: permissive
Topics: knowledge-graph, relation-extraction, chinese, named-entity-recognition, attribute-extraction, low-resource, document-level, information-extraction, pytorch, deepke, ner, nlp, few-shot, prompt, deep-learning, kg, multi-modal, lightner, knowprompt, instructie
Last push: 2026-07-13T16:40:58+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 92, release rhythm 8, longevity 100
- inputs: {"age_days": 2955, "days_push": 51, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4472, forks 750 (observed 2026-08-28T04:08:50.973119+00:00)

## What it is
DeepKE is an open-source PyTorch-based knowledge extraction toolkit for knowledge graph construction, supporting named entity recognition, relation extraction, and attribute extraction. It covers standard supervised, low-resource few-shot, document-level, and multimodal scenarios, and includes LLM-based extraction via DeepKE-LLM and OneKE.

## Use cases
- extract entities and relations from unstructured text to build a knowledge graph
- train a named entity recognition model with few labeled examples
- perform document-level relation extraction across multiple sentences
- run multimodal entity and relation extraction using visual cues
- use large language models for information extraction
- extract relational triples in Chinese with cnSchema support

## When to choose
- building knowledge graphs from text, especially Chinese or low-resource settings
- you need research-grade IE models (NER, RE, triple extraction) with PyTorch
- you want few-shot or multimodal information extraction capabilities
- you want LLM-based extraction with prebuilt OneKE models

## When to avoid
- you need a production-ready REST API or GUI out of the box
- your task is unrelated to information extraction or knowledge graphs
- you need a lightweight pure-inference library without training infrastructure

## Facets
- artifact type: library
- maturity: active
- function: nlp, machine-learning, deep-learning, parser
- domain: machine-learning, artificial-intelligence
- platform: python
- tags: knowledge-graph, information-extraction, named-entity-recognition, relation-extraction, attribute-extraction, low-resource, few-shot, document-level, multimodal, pytorch, chinese-nlp, llm, knowledge-graph-construction, natural-language-processing, linux, gpu

## Member repositories
- zjunlp/DeepKE (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:50.973119+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:20:36.212886+00:00, confidence not recorded.
  - readme: https://github.com/zjunlp/DeepKE (fetched 2026-08-28T04:08:50.973119+00:00, sha f26b422d1cb8)
  - homepage: http://deepke.zjukg.cn/ (fetched 2026-08-29T09:07:06.002126+00:00, sha 5cfa3f609901)
  - registry_pypi: https://pypi.org/pypi/deepke/json (fetched 2026-08-29T09:07:06.011571+00:00, sha 44e7150d07c3)
- Data as of 2026-08-30T08:39:29.467469+00:00.