初始化项目，由ModelHub XC社区提供模型

Model: Elhusseny/DigitalAhmed-V3-qwen2.5-0.5B
Source: Original Platform

README.md

@@ -0,0 +1,41 @@
+---
+tags:
+- gguf
+- qwen
+- qwen2.5
+- llama.cpp
+- text-generation
+- arabic
+- egyptian
+- merged
+- unsloth
+license: apache-2.0
+language:
+- ar
+- en
+---
+
+# DigitalAhmed-V3 (Qwen2.5-0.5B)
+
+"أحمد الرقمي" - نسخة رقمية من شخصية أحمد حسين.
+
+## Model Variants
+
+| File | Quantization | Size |
+|------|-------------|------|
+| ahmed_digital_f16.gguf | F16 | ~1GB |
+| ahmed_digital_q8_0.gguf | Q8_0 | ~0.5GB |
+
+## Usage
+
+```bash
+./llama-cli -m ahmed_digital_q8_0.gguf -p 'أحمد الرقمي هو' -n 100
+```
+
+## Training
+
+- Base Model: Qwen2.5-0.5B-Instruct
+- Method: LoRA Fine-tuning via Unsloth
+- Dataset: Custom Egyptian Arabic instruct
+
+Built with love by Ahmed Hussein.