davidlenz / BERT-SQuAD

SQuAD Question Answering Using BERT, PyTorch

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

BERT-SQuAD

Use google BERT to do SQuAD !

What is SQuAD?

Stanford Question Answering Dataset (SQuAD) is a reading comprehension dataset, consisting of questions posed by crowdworkers on a set of Wikipedia articles, where the answer to every question is a segment of text, or span, from the corresponding reading passage, or the question might be unanswerable.

Requirements

  • python3
  • pip3 install -r requirements.txt

Result

model : bert-large-uncased-whole-word-masking

{"exact_match": 86.91579943235573, "f1": 93.1532499015869}

Pretrained model download from here

unzip and move files to model directory

Inference

from bert import QA

model = QA('model')

doc = "Victoria has a written constitution enacted in 1975, but based on the 1855 colonial constitution, passed by the United Kingdom Parliament as the Victoria Constitution Act 1855, which establishes the Parliament as the state's law-making body for matters coming under state responsibility. The Victorian Constitution can be amended by the Parliament of Victoria, except for certain 'entrenched' provisions that require either an absolute majority in both houses, a three-fifths majority in both houses, or the approval of the Victorian people in a referendum, depending on the provision."

q = 'When did Victoria enact its constitution?'

answer = model.predict(doc,q)

print(answer['answer'])
# 1975
print(ans.keys())
# dict_keys(['answer', 'start', 'end', 'confidence', 'document']))

model.predict(doc,q) return dict

{
"answer" : "answer text",
"start" : "start index",
"end" : "end index",
"confiednce" : "confidence of answer",
"document" : "tokenzied document , list"
}

Deploy REST-API

BERT QA model deployed as rest api

python api.py

API will be live at 0.0.0.0:8000 endpoint predict

cURL request

curl -X POST http://0.0.0.0:8000/predict -H 'Content-Type: application/json' -d '{ "document": "Victoria has a written constitution enacted in 1975, but based on the 1855 colonial constitution, passed by the United Kingdom Parliament as the Victoria Constitution Act 1855, which establishes the Parliament as the states law-making body for matters coming under state responsibility. The Victorian Constitution can be amended by the Parliament of Victoria, except for certain 'entrenched' provisions that require either an absolute majority in both houses, a three-fifths majority in both houses, or the approval of the Victorian people in a referendum, depending on the provision.","question":"When did Victoria enact its constitution?" }'

Output

{
    "result": {
        "answer": "1975",
        "confidence": 0.940746070672879,
        "document": [
            "Victoria",
            "has",
            "a",
            "written",
            "constitution",
            "enacted",
            "in",
            "1975,",
            "but",
            "based",
            "on",
            "the",
            "1855",
            "colonial",
            "constitution,",
            "passed",
            "by",
            "the",
            "United",
            "Kingdom",
            "Parliament",
            "as",
            "the",
            "Victoria",
            "Constitution",
            "Act",
            "1855,",
            "which",
            "establishes",
            "the",
            "Parliament",
            "as",
            "the",
            "states",
            "law-making",
            "body",
            "for",
            "matters",
            "coming",
            "under",
            "state",
            "responsibility.",
            "The",
            "Victorian",
            "Constitution",
            "can",
            "be",
            "amended",
            "by",
            "the",
            "Parliament",
            "of",
            "Victoria,",
            "except",
            "for",
            "certain",
            "entrenched",
            "provisions",
            "that",
            "require",
            "either",
            "an",
            "absolute",
            "majority",
            "in",
            "both",
            "houses,",
            "a",
            "three-fifths",
            "majority",
            "in",
            "both",
            "houses,",
            "or",
            "the",
            "approval",
            "of",
            "the",
            "Victorian",
            "people",
            "in",
            "a",
            "referendum,",
            "depending",
            "on",
            "the",
            "provision."
        ],
        "end": 7,
        "start": 7
    }
}

Postman

postman output image

About

SQuAD Question Answering Using BERT, PyTorch

License:GNU Affero General Public License v3.0


Languages

Language:Python 100.0%