AIDC-LAOS/aidc-llm-laos-4b
AIDC-LAOS/aidc-llm-laos-4b is a 4.3 billion parameter language model fine-tuned by Phonepadith based on Google's Gemma 3-4B Instruct architecture. This model is specifically adapted for the Lao language, excelling in summarization, question answering, and general chat tasks. It was trained on 24,000 high-quality Lao input-output pairs, making it suitable for generating concise summaries from Lao text, including news, government documents, and meeting minutes.
Loading preview...
AIDC-LAOS/aidc-llm-laos-4b: Lao Language Model
This model is a fine-tuned version of Google's Gemma 3-4B Instruct, specifically adapted for the Lao language (๐ฑ๐ฆ). Developed by Phonepadith, it leverages a curated dataset of over 24,000 high-quality Lao input-output pairs, with a focus on AIDC (Artificial Intelligence and Digital Community) topics.
Key Capabilities
- Lao Language Generation: Optimized to understand and generate responses in Lao.
- Summarization: Designed to create concise summaries from various Lao texts, such as news, government documents, and meeting minutes.
- General Chat & Question Answering: Capable of engaging in general conversations and answering questions in Lao.
Training Details
The model was fine-tuned using LoRA with PEFT (Unsloth) on a sequence length of 2048. It utilized an AdamW optimizer over 3-5 epochs with early stopping enabled. The base model is google/gemma-3-4b-it, and the fine-tuned model is available in GGUF (F16, Q8_0) format.
Good For
- Applications requiring Lao text summarization.
- Developing Lao-speaking chatbots or virtual assistants.
- Projects focused on Lao language understanding and generation.