T

TC Instruct DPO

Developed by tanamettpk
Thai instruction-optimized model fine-tuned from Typhoon-7B using Direct Preference Optimization (DPO) technology
Downloads 28
Release Time : 2/17/2024

Model Overview

This model is a Thai instruction-optimized model fine-tuned from SCB 10X's Typhoon-7B (derived from Mistral-7B), specifically developed for researching large language model construction processes. Trained using QLoRA technology, it supports various Thai instruction tasks.

Model Features

Thai Instruction Optimization
Specifically optimized for Thai instructions to ensure instruction diversity
Direct Preference Optimization (DPO)
Trained using Direct Preference Optimization technology to improve response quality
QLoRA Efficient Fine-tuning
Efficient fine-tuning using QLoRA technology (rank 32, alpha value 64)

Model Capabilities

Thai text generation
Instruction following
Q&A system

Use Cases

Research Applications
Large Language Model Construction Research
Used for researching Thai large language model construction processes and techniques
Dialogue Systems
Thai Chatbot
Can be used to build Thai dialogue systems
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase