toandev/donglao-gemma-3-4b-it-vi

Overview

The toandev/donglao-gemma-3-4b-it-vi model is a specialized large language model (LLM) built upon Google's Gemma 3 4B Instruct architecture. With 4.3 billion parameters and a substantial context length of 32768 tokens, this model is primarily distinguished by its fine-tuning for the Vietnamese language.

Key Capabilities

• Vietnamese Language Proficiency: Optimized for understanding and generating text in Vietnamese.
• Instruction Following: Inherits instruction-tuned capabilities from its base Gemma 3 4B Instruct model.
• Multimodal Data Training: Fine-tuned using the 5CD-AI/Viet-ShareGPT-4o-Text-VQA dataset, which includes Vietnamese text and visual question answering data.

Good For

• Vietnamese Text Generation: Creating coherent and contextually relevant Vietnamese text.
• Vietnamese Question Answering: Responding to queries in Vietnamese, potentially leveraging its training on VQA data.
• Vietnamese Conversational AI: Developing chatbots or conversational agents that interact effectively in Vietnamese.