SILMA-9B-Instruct-v1.0-GGUF
| Property | Value |
|---|---|
| Parameter Count | 9.24B |
| License | Gemma |
| Languages | Arabic, English |
| Format | GGUF (Various Quantizations) |
What is SILMA-9B-Instruct-v1.0-GGUF?
SILMA-9B-Instruct-v1.0-GGUF is a bilingual large language model optimized for Arabic and English language processing, available in multiple GGUF quantizations for efficient deployment. The model demonstrates strong performance on Arabic language benchmarks, achieving 52.55% on MMLU (Arabic) and 71.85% on AlGhafa benchmark.
Implementation Details
The model is available in various quantization formats ranging from 3.544GB (Q2_K) to 9.152GB (Q8_0), offering different trade-offs between model size and quality. The recommended variants are Q4_K_M (5.365GB) for balanced performance and Q5_K_M (6.191GB) for higher quality.
- Supports custom prompt template with system prompts
- Compatible with llama.cpp framework
- Multiple quantization options for different deployment scenarios
Core Capabilities
- Bilingual processing in Arabic and English
- Strong performance on Arabic language benchmarks
- Efficient deployment through GGUF format
- Flexible quantization options for different hardware constraints
Frequently Asked Questions
Q: What makes this model unique?
SILMA-9B stands out for its specialized optimization for Arabic language processing while maintaining English capabilities, offering various quantization options for efficient deployment across different hardware configurations.
Q: What are the recommended use cases?
The model is well-suited for Arabic-English bilingual applications, particularly in scenarios requiring efficient deployment through GGUF format. The different quantization options allow for deployment on various hardware configurations, from resource-constrained to high-performance environments.
