QNLI-Electra-Base Open-Source Model - Free Assistance for Question Answering NLI, Accurately Judge Whether a Paragraph Can Answer Questions

Qnli Electra Base

Developed by cross-encoder

This is a cross-encoder model based on the ELECTRA architecture, specifically designed for natural language inference (NLI) in question-answering tasks, determining whether a given question can be answered by a specific paragraph.

Question Answering System

Transformers

EnglishOpen Source License:Apache-2.0 #Question-Answer Matching #Natural Language Inference #Paragraph Relevance Scoring

Downloads 6,172

Release Time : 3/2/2022

Model Overview

This model is trained using the SentenceTransformers framework to evaluate the relevance between questions and paragraphs, determining whether the paragraph can answer the question.

Model Features

Based on ELECTRA Architecture

Uses the efficient ELECTRA-base-discriminator as the base model.

Cross-Encoder Design

Specifically designed for paired text input (question-paragraph) scoring tasks.

SQuAD/QNLI Training

Trained on the GLUE QNLI dataset, which is derived from the SQuAD question-answering dataset.

Model Capabilities

Question-Paragraph Relevance Scoring

Natural Language Inference

Question-Answering System Support

Use Cases

Question-Answering Systems

Question-Answer Relevance Judgment

Evaluate the relevance between a user's question and candidate answer paragraphs.

Provides a relevance score between 0 and 1.

Information Retrieval

Document Paragraph Ranking

Rank retrieved document paragraphs by relevance.

Helps filter the paragraphs most likely to contain the answer.

🚀 Cross-Encoder for SQuAD (QNLI)

This model is designed for text ranking, trained to determine if a question can be answered by a given paragraph.

Property	Details
Model Type	Cross-Encoder for SQuAD (QNLI)
Base Model	google/electra-base-discriminator
Pipeline Tag	text-ranking
Library Name	sentence-transformers
Tags	transformers

🚀 Quick Start

This model was trained using SentenceTransformers Cross-Encoder class.

📦 Installation

No specific installation steps are provided in the original README.

💻 Usage Examples

Basic Usage

Pre-trained models can be used like this:

from sentence_transformers import CrossEncoder

model = CrossEncoder('cross-encoder/qnli-electra-base')
scores = model.predict([('Query1', 'Paragraph1'), ('Query2', 'Paragraph2')])

#e.g.
scores = model.predict([('How many people live in Berlin?', 'Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers.'), ('What is the size of New York?', 'New York City is famous for the Metropolitan Museum of Art.')])

Advanced Usage

You can use the model also directly with Transformers library (without SentenceTransformers library):

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model = AutoModelForSequenceClassification.from_pretrained('cross-encoder/qnli-electra-base')
tokenizer = AutoTokenizer.from_pretrained('cross-encoder/qnli-electra-base')

features = tokenizer(['How many people live in Berlin?', 'What is the size of New York?'], ['Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers.', 'New York City is famous for the Metropolitan Museum of Art.'],  padding=True, truncation=True, return_tensors="pt")

model.eval()
with torch.no_grad():
    scores = torch.nn.functional.sigmoid(model(**features).logits)
    print(scores)

📚 Documentation

Training Data

Given a question and paragraph, can the question be answered by the paragraph? The models have been trained on the GLUE QNLI dataset, which transformed the SQuAD dataset into an NLI task.

Performance

For performance results of this model, see SBERT.net Pre-trained Cross-Encoder.

📄 License

This project is licensed under the Apache-2.0 license.

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