| Property | Value |
|---|---|
| Parameter Count | 226M |
| Base Model | VietAI/vit5-base |
| License | MIT |
| Language | Vietnamese |
| Task | Abstractive Question Answering |
What is ViBidLAQA_base?
ViBidLAQA_base is a specialized Vietnamese language model designed for abstractive question answering in the bidding law domain. Built on the VietAI/vit5-base architecture, this model represents a significant advancement in Vietnamese legal AI applications, offering precise and natural responses to bidding law queries.
Implementation Details
The model is trained on a carefully curated dataset comprising 5,300 training samples and 1,000 test samples. The training data was generated using Claude 3.5 Sonnet and validated by legal experts, while the test set was manually created by Vietnamese legal professionals. The model demonstrates impressive performance metrics, including a ROUGE-1 score of 75.09 and a BERT-Score of 86.65.
- Fine-tuned T5 architecture optimized for Vietnamese legal text
- Implements sequence-to-sequence learning for natural answer generation
- Supports context-aware question answering with bidding law expertise
Core Capabilities
- Generates natural, contextually accurate responses to bidding law questions
- Processes both questions and context efficiently with a 512-token input limit
- Achieves high accuracy in legal domain question answering
- Supports integration with legal information systems and chatbots
Frequently Asked Questions
Q: What makes this model unique?
This model is specifically optimized for Vietnamese bidding law, combining domain expertise with advanced language modeling capabilities. Its specialized training on legal content and expert-validated data sets it apart from general-purpose language models.
Q: What are the recommended use cases?
The model is ideal for building legal advisory systems, automating bidding law query responses, and supporting legal professionals in information retrieval. It's particularly useful for organizations dealing with Vietnamese bidding procedures and regulations.