airesearch/WangchanX-Legal-ThaiCCL-Retriever

WangchanX-Legal-ThaiCCL-Retriever

This model card describes WangchanX-Legal-ThaiCCL-Retriever, a retriever model fine-tuned from the bge-m3 model on the WangchanX-Legal-ThaiCCL-RAG dataset. It is designed to retrieve relevant legal text sections in response to legal questions posed in Thai, specifically focusing on Corporate and Commercial Law (CCL).

Model Details:

• Base Model: bge-m3
• Fine-tuned Dataset: WangchanX-Legal-ThaiCCL-RAG dataset
• Language: Thai
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 1024 tokens
• License: MIT

WangchanX-Legal-ThaiCCL-RAG

This dataset focuses on supporting Thai legal question-answering systems using Retrieval-Augmented Generation (RAG), focusing on Corporate and Commercial Law.

Intended Use Cases: This model is designed for use as a retriever model within a larger RAG pipeline.

• Legal Question Answering: Serving as a core component in a larger question-answering system that provides answers to user queries about Thai law.
• Legal Information Retrieval: Enabling efficient retrieval of information from Thai legal texts.