Gemma2 9B CPT Sahabat-AI v1 Instruct GGUF
| Property | Value |
|---|---|
| Parameter Count | 9.24B |
| Languages | English, Indonesian, Javanese, Sundanese |
| License | Gemma Community License |
| Author | QuantFactory |
| Base Model | GoToCompany/gemma2-9b-cpt-sahabatai-v1-base |
What is gemma2-9b-cpt-sahabatai-v1-instruct-GGUF?
This is a quantized version of the Sahabat-AI language model, specifically designed for Indonesian language and its dialects. The model has been fine-tuned on 448,000 Indonesian instruction-completion pairs, 96,000 Javanese pairs, 98,000 Sundanese pairs, and 129,000 English pairs, making it particularly effective for multilingual applications in the Indonesian context.
Implementation Details
The model is built on the Gemma2 architecture with a context length of 8192 tokens. It utilizes GGUF quantization for efficient deployment and has shown superior performance in benchmarks compared to other models in its class.
- Achieves 61.169% overall score on SEA HELM benchmark
- Scores 62.6% on IndoMMLU evaluation
- Demonstrates strong performance in English tasks with 33.67% average score
Core Capabilities
- Multilingual understanding and generation across four languages
- Strong instruction-following capabilities
- Advanced question-answering and sentiment analysis
- Effective translation and summarization
- Cultural context awareness for Indonesian dialects
Frequently Asked Questions
Q: What makes this model unique?
The model's specialization in Indonesian and its dialects, combined with its impressive performance on regional benchmarks, makes it particularly valuable for applications targeting Indonesian users. Its multilingual capabilities and cultural awareness set it apart from general-purpose LLMs.
Q: What are the recommended use cases?
The model excels in tasks requiring understanding of Indonesian culture and language nuances, including content generation, translation, educational applications, and customer service automation for Indonesian-speaking regions.
