SpanBERT Open-Source Question Answering Model - Finetuned on SQuAD 2.0, Freely Implement High-Efficiency Q&A Applications

Spanbert Finetuned Squadv2

Developed by mrm8488

SpanBERT is a pre-trained model developed by Facebook Research, fine-tuned on the SQuAD 2.0 dataset for question answering tasks.

Question Answering System English#Question Answering System #Text Span Prediction #SQuAD2.0 Adaptation

Downloads 154

Release Time : 3/2/2022

Model Overview

This model is a question-answering model based on the SpanBERT architecture, specifically fine-tuned for the SQuAD 2.0 dataset, capable of answering questions and determining whether a question is answerable.

Model Features

Text Span Prediction

SpanBERT improves pre-training by representing and predicting text spans, making it particularly suitable for question-answering tasks.

Handling Unanswerable Questions

The model can identify whether a passage does not support an answer and avoid responding, which is a core challenge of SQuAD 2.0.

Model Capabilities

Question Answering System

Text Understanding

Answer Localization

Use Cases

Education

Reading Comprehension Assistance

Helps students understand text content and answer questions

Improves reading comprehension efficiency and accuracy

Customer Service

Automated Q&A System

Automatically answers common customer questions

Reduces workload for manual customer support

🚀 SpanBERT (spanbert-base-cased) fine-tuned on SQuAD v2

SpanBERT, developed by Facebook Research, is fine-tuned on SQuAD 2.0 for the Q&A downstream task, offering high - performance question - answering capabilities.

🚀 Quick Start

SpanBERT, created by Facebook Research, is fine-tuned on SQuAD 2.0 for the Q&A downstream task.

✨ Features

Details of SpanBERT

SpanBERT: Improving Pre-training by Representing and Predicting Spans

Details of the downstream task (Q&A) - Dataset

SQuAD2.0 combines the 100,000 questions in SQuAD1.1 with over 50,000 unanswerable questions written adversarially by crowdworkers to look similar to answerable ones. To do well on SQuAD2.0, systems must not only answer questions when possible, but also determine when no answer is supported by the paragraph and abstain from answering.

Dataset	Split	# samples
SQuAD2.0	train	130k
SQuAD2.0	eval	12.3k

🔧 Technical Details

Model training

The model was trained on a Tesla P100 GPU and 25GB of RAM. The script for fine - tuning can be found here

📚 Documentation

Results

Metric	# Value
EM	78.80
F1	82.22

Raw metrics

{
  "exact": 78.80064010780762,
  "f1": 82.22801347271162,
  "total": 11873,
  "HasAns_exact": 78.74493927125506,
  "HasAns_f1": 85.60951483831069,
  "HasAns_total": 5928,
  "NoAns_exact": 78.85618166526493,
  "NoAns_f1": 78.85618166526493,
  "NoAns_total": 5945,
  "best_exact": 78.80064010780762,
  "best_exact_thresh": 0.0,
  "best_f1": 82.2280134727116,
  "best_f1_thresh": 0.0
}

Comparison

Model	EM	F1 score
SpanBert official repo	-	83.6*
spanbert-finetuned-squadv2	78.80	82.22

💻 Usage Examples

Basic Usage

from transformers import pipeline

qa_pipeline = pipeline(
    "question-answering",
    model="mrm8488/spanbert-finetuned-squadv2",
    tokenizer="mrm8488/spanbert-finetuned-squadv2"
)

qa_pipeline({
    'context': "Manuel Romero has been working hardly in the repository hugginface/transformers lately",
    'question': "Who has been working hard for hugginface/transformers lately?"

})

# Output: {'answer': 'Manuel Romero','end': 13,'score': 6.836378586818937e-09, 'start': 0}

Created by Manuel Romero/@mrm8488

Made with ♥ in Spain

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご