V

Vi Qwen2 7B RAG

Developed by AITeamVN
Vietnamese RAG-specific large language model fine-tuned from Qwen2-7B-Instruct, supports 8192-token context length
Downloads 737
Release Time : 10/1/2024

Model Overview

Model optimized for Vietnamese Retrieval-Augmented Generation (RAG) tasks, featuring core capabilities such as noise resistance, negative example filtering, and multi-document integration

Model Features

Noise Resistance
Can accurately extract valid information from noisy documents (supports 1 positive + 4 negative examples scenario)
Negative Example Filtering
Can refuse to answer when retrieved documents don't contain the answer (supports 1-6 negative examples scenario)
Multi-document Integration
Can synthesize information from 2-3 partially relevant documents to answer complex questions
High Accuracy Judgment
99% accuracy in determining whether the context contains the answer

Model Capabilities

Vietnamese text generation
Retrieval-Augmented Generation
Multi-document information integration
Context relevance judgment
Noise-resistant processing
Negative example filtering

Use Cases

Medical Information Processing
Drug Manual Parsing
Extract indications and dosage instructions from complex drug manuals
Example shows complete extraction of Insuact 10mg medication information
Historical Data Analysis
Historical Event Analysis
Integrate key information from multiple historical materials
Successfully analyzed three main reasons why Qing Dynasty princes didn't rebel
Legal Procedure Queries
Judicial Process Explanation
Answer specific procedural questions by synthesizing multiple legal documents
Accurately explained the handling process for unclaimed bodies after lethal injection execution
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase