AIbase
EN
Home
Viet Llama2 Ft
V

Viet Llama2 Ft

Developed by infCapital
A Vietnamese-optimized model based on LLaMa2-7B, with expanded vocabulary and continuous pre-training on 2 billion Vietnamese tokens, supporting Vietnamese comprehension and generation tasks.
Large Language Model
Transformers
 Other#Vietnamese optimization#Multi-source instruction fine-tuning#Vocabulary expansion
Downloads 92
Release Time : 9/28/2023

Model Overview

This model is a Vietnamese-optimized version of LLaMa2-7B, optimized for Vietnamese comprehension and generation tasks through vocabulary expansion and extensive Vietnamese data pre-training.

Model Features

Vietnamese optimization
Expanded vocabulary to 44,800 to better support Vietnamese and continuously pre-trained on 2 billion Vietnamese tokens
Multi-source data fine-tuning
Fine-tuned based on a Vietnamese dataset integrated from multiple sources and translated via GPT-3
Large model foundation
Based on the powerful LLaMa2-7B model architecture, with excellent language comprehension and generation capabilities

Model Capabilities

Vietnamese text generation
Vietnamese dialogue systems
Vietnamese language understanding
Multi-turn dialogue
Instruction following

Use Cases

Dialogue systems
Vietnamese customer service chatbot
Dialogue system for Vietnamese customer service
Capable of understanding and generating fluent Vietnamese customer service conversations
Content generation
Vietnamese content creation
Generating Vietnamese articles, stories, and other content
Can generate fluent text that conforms to Vietnamese language conventions
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase