初始化项目，由ModelHub XC社区提供模型

Model: NightPrince/Qwen3-4B-Islamic-Arabic-GGUF Source: Original Platform
2026-05-25 16:16:16 +08:00
commit 8361dfc96a
5 changed files with 242 additions and 0 deletions
--- a/.gitattributes
+++ b/.gitattributes
@@ -0,0 +1,38 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+qwen3-4b-islamic-f16.gguf filter=lfs diff=lfs merge=lfs -text
+qwen3-4b-islamic-q4_k_m.gguf filter=lfs diff=lfs merge=lfs -text
+qwen3-4b-islamic-q8_0.gguf filter=lfs diff=lfs merge=lfs -text
--- a/README.md
+++ b/README.md
@@ -0,0 +1,195 @@
+---
+language:
+  - ar
+license: apache-2.0
+library_name: gguf
+base_model: NightPrince/Qwen3-4B-Islamic-Arabic
+datasets:
+  - NightPrince/islamic-arabic-qa
+tags:
+  - arabic
+  - islamic
+  - fiqh
+  - fatwa
+  - qwen3
+  - gguf
+  - llama-cpp
+  - ollama
+  - quantized
+  - instruction-tuning
+pipeline_tag: text-generation
+---
+
+# Qwen3-4B-Islamic-Arabic-GGUF
+
+**GGUF quantized versions of Qwen3-4B-Islamic-Arabic for llama.cpp, Ollama, and LM Studio.**
+
+This repository contains three GGUF files at different quantization levels, converted from [NightPrince/Qwen3-4B-Islamic-Arabic](https://huggingface.co/NightPrince/Qwen3-4B-Islamic-Arabic) (the merged FP16 model). All standard GGUF-compatible runtimes are supported: llama.cpp, Ollama, LM Studio, Jan, and others.
+
+Trained and converted by **[Yahya Alnwsany (NightPrince)](https://huggingface.co/NightPrince)** — 2026-05-05.
+
+---
+
+## Files
+
+| File | Size | Recommended for |
+|---|---|---|
+| `qwen3-4b-islamic-q4_k_m.gguf` | 2.3 GB | **Most users** — best quality/size balance |
+| `qwen3-4b-islamic-q8_0.gguf` | 4.0 GB | High quality, more RAM available |
+| `qwen3-4b-islamic-f16.gguf` | 7.5 GB | Reference / re-quantization source |
+
+**Recommendation**: Start with `q4_k_m`. If you have 6+ GB of RAM headroom and want noticeably sharper Arabic output, use `q8_0`. The `f16` file is the lossless reference and is best used as a source for producing custom quantizations with llama.cpp's `llama-quantize`.
+
+---
+
+## Model Variants
+
+| Variant | Repo | Description |
+|---|---|---|
+| **Merged FP16** | [NightPrince/Qwen3-4B-Islamic-Arabic](https://huggingface.co/NightPrince/Qwen3-4B-Islamic-Arabic) | Canonical merged model, FP16, ~7.6 GB — drop-in for transformers or vLLM |
+| **LoRA Adapter** | [NightPrince/Qwen3-4B-Islamic-Arabic-LoRA](https://huggingface.co/NightPrince/Qwen3-4B-Islamic-Arabic-LoRA) | PEFT adapter only, 264 MB — apply on top of `Qwen/Qwen3-4B` |
+| **INT4 Quantized** | [NightPrince/Qwen3-4B-Islamic-Arabic-INT4](https://huggingface.co/NightPrince/Qwen3-4B-Islamic-Arabic-INT4) | W4A16 compressed-tensors for fast vLLM serving, 2.5 GB |
+| **MLX 4-bit** | [NightPrince/Qwen3-4B-Islamic-Arabic-mlx-4Bit](https://huggingface.co/NightPrince/Qwen3-4B-Islamic-Arabic-mlx-4Bit) | Apple Silicon / MLX — native Mac inference, 4-bit quantized |
+| **GGUF** (this model) | [NightPrince/Qwen3-4B-Islamic-Arabic-GGUF](https://huggingface.co/NightPrince/Qwen3-4B-Islamic-Arabic-GGUF) | llama.cpp / Ollama / LM Studio — Q4_K_M (2.3 GB), Q8_0 (4.0 GB), F16 (7.5 GB) |
+| **Dataset** | [NightPrince/islamic-arabic-qa](https://huggingface.co/datasets/NightPrince/islamic-arabic-qa) | 17,944 train / 2,101 val / 1,042 test — Islamic Arabic Q&A pairs |
+
+---
+
+## Usage
+
+### Ollama
+
+**Step 1: Create a Modelfile**
+
+Save the following as `Modelfile` (no extension) in any directory:
+
+```
+FROM ./qwen3-4b-islamic-q4_k_m.gguf
+
+SYSTEM """أنت مساعد عالم إسلامي متخصص. أجب على الأسئلة بدقة استناداً إلى القرآن الكريم والسنة النبوية والفقه الإسلامي الكلاسيكي. استشهد بالمصادر حيثما أمكن. كن موجزاً لكن شاملاً."""
+
+PARAMETER temperature 0.7
+PARAMETER top_p 0.9
+PARAMETER num_ctx 4096
+```
+
+> Important: The `SYSTEM` field above contains the exact system prompt the model was fine-tuned with. Using it will produce the best results.
+
+**Step 2: Download the GGUF file**
+
+```bash
+# Using huggingface-cli
+pip install huggingface_hub
+huggingface-cli download NightPrince/Qwen3-4B-Islamic-Arabic-GGUF \
+    qwen3-4b-islamic-q4_k_m.gguf \
+    --local-dir .
+```
+
+**Step 3: Build and run**
+
+```bash
+# Create the Ollama model
+ollama create qwen3-islamic -f Modelfile
+
+# Run interactively
+ollama run qwen3-islamic
+
+# Or query via API
+curl http://localhost:11434/api/generate -d '{
+  "model": "qwen3-islamic",
+  "prompt": "ما حكم الاحتفال بالمولد النبوي الشريف؟",
+  "stream": false
+}'
+```
+
+---
+
+### llama.cpp
+
+**Build llama.cpp** (if not already installed):
+
+```bash
+git clone https://github.com/ggerganov/llama.cpp
+cd llama.cpp
+make -j$(nproc)            # CPU
+# For CUDA: make GGML_CUDA=1 -j$(nproc)
+```
+
+**Download a GGUF file:**
+
+```bash
+huggingface-cli download NightPrince/Qwen3-4B-Islamic-Arabic-GGUF \
+    qwen3-4b-islamic-q4_k_m.gguf \
+    --local-dir ./models
+```
+
+**Run the llama.cpp HTTP server:**
+
+```bash
+./llama-server \
+    --model ./models/qwen3-4b-islamic-q4_k_m.gguf \
+    --ctx-size 4096 \
+    --n-gpu-layers 99 \
+    --host 0.0.0.0 \
+    --port 8080 \
+    --system-prompt "أنت مساعد عالم إسلامي متخصص. أجب على الأسئلة بدقة استناداً إلى القرآن الكريم والسنة النبوية والفقه الإسلامي الكلاسيكي. استشهد بالمصادر حيثما أمكن. كن موجزاً لكن شاملاً."
+```
+
+**CLI inference:**
+
+```bash
+./llama-cli \
+    --model ./models/qwen3-4b-islamic-q4_k_m.gguf \
+    --ctx-size 4096 \
+    --n-gpu-layers 99 \
+    --chat-template qwen3 \
+    --system-prompt "أنت مساعد عالم إسلامي متخصص. أجب على الأسئلة بدقة استناداً إلى القرآن الكريم والسنة النبوية والفقه الإسلامي الكلاسيكي. استشهد بالمصادر حيثما أمكن. كن موجزاً لكن شاملاً." \
+    --prompt "ما هي أركان الإسلام الخمسة؟" \
+    --n-predict 512
+```
+
+---
+
+### LM Studio
+
+1. Open **LM Studio** and go to the **Search** tab.
+2. Search for `NightPrince/Qwen3-4B-Islamic-Arabic-GGUF`.
+3. Download `qwen3-4b-islamic-q4_k_m.gguf` (recommended) from the file list.
+4. Load the model and open the **Chat** tab.
+5. In **System Prompt**, paste:
+   ```
+   أنت مساعد عالم إسلامي متخصص. أجب على الأسئلة بدقة استناداً إلى القرآن الكريم والسنة النبوية والفقه الإسلامي الكلاسيكي. استشهد بالمصادر حيثما أمكن. كن موجزاً لكن شاملاً.
+   ```
+6. Set **Temperature** to `0.7` and **Context Length** to `4096` for best results.
+
+---
+
+## Hardware Requirements
+
+| File | Min RAM (CPU) | Min VRAM (GPU offload) |
+|---|---|---|
+| `q4_k_m` (2.3 GB) | 4 GB | 3–4 GB |
+| `q8_0` (4.0 GB) | 6 GB | 5–6 GB |
+| `f16` (7.5 GB) | 10 GB | 8–10 GB |
+
+Use `--n-gpu-layers 99` in llama.cpp to offload all layers to GPU. Reduce the value if you run out of VRAM.
+
+---
+
+## Citation
+
+```bibtex
+@misc{alnwsany2026qwen3islamicarbic,
+  author       = {Yahya Alnwsany},
+  title        = {Qwen3-4B-Islamic-Arabic: QLoRA Fine-Tuning of Qwen3-4B on Islamic Arabic Q\&A},
+  year         = {2026},
+  howpublished = {\url{https://huggingface.co/NightPrince/Qwen3-4B-Islamic-Arabic}},
+  note         = {Base model: Qwen/Qwen3-4B. Dataset: NightPrince/islamic-arabic-qa.}
+}
+```
+
+---
+
+## License
+
+Apache 2.0 — consistent with the base model [Qwen/Qwen3-4B](https://huggingface.co/Qwen/Qwen3-4B).
--- a/qwen3-4b-islamic-f16.gguf
+++ b/qwen3-4b-islamic-f16.gguf
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:894aab77eb0f938aa07078bca49ed36cc1df1c956626c02552c8bf004ca15bba
+size 8051284800
--- a/qwen3-4b-islamic-q4_k_m.gguf
+++ b/qwen3-4b-islamic-q4_k_m.gguf
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d907f52954a81597f0b744e7ce3c9dea9778c116fded0f4e8c3f39bf45b7834
+size 2497280320
--- a/qwen3-4b-islamic-q8_0.gguf
+++ b/qwen3-4b-islamic-q8_0.gguf
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d11109087fbfc6b07bee06e1884d6a71a906c2d56131a659d9f5f1c657ccaa0
+size 4280404800