Sign inSign up

uhhlt/maverick-de

By uhhlt

•Updated 7 months ago

German Maverick Coref

Image
0

1.1K

uhhlt/maverick-de repository overview

⁠German Maverick Coref

German coreference resolution using https://github.com/uhh-lt/maverick-coref-de⁠

Run container, e.g. docker run --rm --name maverick-de -p 8080:8080 --gpus '"device=0"' uhhlt/maverick-de:latest

It supports two environment parameters: MODEL_NAME (to use another model than the default) and DEVICE (defaults to cuda, you can set cpu to try it without a GPU...)

Support plain, tokenized, and tokenized+sentence split (recommended) text:

curl -X 'POST' \
  'http://localhost:8080/predict' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "tokens": 
      "Alice verkauft ihr altes Fahrrad. Sie braucht es nicht mehr."
}'

or

curl -X 'POST' \
  'http://localhost:8080/predict' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "tokens": 
      ["Alice",  "verkauft", "ihr", "altes", "Fahrrad","." ,"Sie", "braucht", "es", "nicht", "mehr", "."]
}'

or

curl -X 'POST' \
  'http://localhost:8080/predict' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
  "tokens": 
      [["Alice",  "verkauft", "ihr", "altes", "Fahrrad","."],[ "Sie", "braucht", "es", "nicht", "mehr", "."]]
}'

Result:

{
  "tokens": [
    "Alice",
    "verkauft",
    "ihr",
    "altes",
    "Fahrrad",
    ".",
    "Sie",
    "braucht",
    "es",
    "nicht",
    "mehr",
    "."
  ],
  "clusters_token_offsets": [
    [
      [
        0,
        0
      ],
      [
        2,
        2
      ],
      [
        6,
        6
      ]
    ],
    [
      [
        2,
        4
      ],
      [
        8,
        8
      ]
    ]
  ],
  "clusters_char_offsets": null,
  "clusters_token_text": [
    [
      "Alice",
      "ihr",
      "Sie"
    ],
    [
      "ihr altes Fahrrad",
      "es"
    ]
  ],
  "clusters_char_text": null
}

⁠Citation

If you use this software, please consider citing our paper published at KONVENS 2025:

@inproceedings{petersenfrey-etal-2025-efficient,
    title = "Efficient and effective coreference resolution for German",
    author = "Petersen-Frey, Fynn and Hatzel, Hans Ole and Biemann, Chris",
    booktitle = "Proceedings of the 21st Conference on Natural Language Processing (KONVENS 2025). Volume 1: Long and Short Papers",
    month = "9",
    year = "2025",
    address = "Hildesheim, Germany",
    publisher = "KONVENS 2025 Organizers"
}

Tag summary

Content type

Image

Digest

sha256:c21f9d69c…

Size

4.2 GB

Last updated

7 months ago

docker pull uhhlt/maverick-de