Skip to content

Devininthelab/open-domain-QA

Repository files navigation

Open Domain Question Answering

Pretrained

Prerequisites

datasets==2.16.1
evaluate==0.4.1
transformers[sentencepiece]==4.35.2
accelerate==0.26.1

Inference

# Use a pipeline as a high-level helper
from transformers import pipeline

PIPELINE_NAME = 'question-answering'
MODEL_NAME = 'winvswon78/distilbert-finetuned-squadv2'
pipe = pipeline(PIPELINE_NAME, model=MODEL_NAME)

#Example
INPUT_QUESTION = 'Where is the cat?'
INPUT_CONTEXT = 'The cat sits under the table, trying to catch a mouse.'
pipe(question=INPUT_QUESTION, context=INPUT_CONTEXT)

Metrics

Metric Score
Exact 49.77
F1 53.75
Total 11873
HasAns Exact 75.52
HasAns F1 83.49
HasAns Total 5928
NoAns Exact 24.10
NoAns F1 24.10
NoAns Total 5945
Best Exact 65.27
Best Exact Thresh -11.46
Best F1 67.29
Best F1 Thresh -9.32

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published