File size: 5,695 Bytes

---
license: apache-2.0
language:
  - en
base_model: Qwen/Qwen2.5-3B-Instruct
tags:
  - homedock
  - homedock-os
  - self-hosting
  - docker
  - home-server
  - documentation
  - qwen2.5
  - gguf
  - fine-tuned
library_name: transformers
pipeline_tag: text-generation
model-index:
  - name: HomeDock-OS-3B-Instruct
    results: []
---

<img src="https://www.homedock.cloud/images/ctxs/shiny-big-960w.webp" alt="HomeDock OS" width="full" align="center">

# HomeDock-OS-3B-Instruct

A fine-tuned version of **Qwen 2.5 3B Instruct** specialized in [HomeDock OS](https://www.homedock.cloud) documentation, features, for guidance and to make self-hosting easier.

## Model Description

HomeDock-OS-3B-Instruct is trained on the complete HomeDock OS knowledge base:

- **Documentation**: Installation, configuration, SSL, troubleshooting
- **200+ Applications Features**: Plex, Jellyfin, Nextcloud, Home Assistant, and more
- **Blog Articles**: Tutorials, comparisons, and best practices
- **Enterprise Features**: Data Space, Secure Tunnel, Audit Trail, and more

### Key Features

- Answers questions about HomeDock OS features and configuration
- Provides guidance on self-hosted application setup
- Understands Docker containerization concepts
- Explains App Store applications and their use cases

## Quantization

| File | Size | Description |
|------|------|-------------|
| `HomeDock-OS-3B-Instruct.Q4_K_M.gguf` | ~1.8 GB | Recommended, best balance |
| `HomeDock-OS-3B-Instruct.Q8_0.gguf` | ~3.2 GB | Higher quality |

## Usage

### Ollama

1. Create a `Modelfile`:

```
FROM ./HomeDock-OS-3B-Instruct.Q4_K_M.gguf

TEMPLATE """<|im_start|>system
{{ .System }}<|im_end|>
<|im_start|>user
{{ .Prompt }}<|im_end|>
<|im_start|>assistant
"""

SYSTEM """You are an expert assistant for HomeDock OS documentation. Provide clear, accurate, and helpful information about HomeDock OS features, configuration, and usage."""

PARAMETER stop "<|im_end|>"
PARAMETER stop "<|im_start|>"
PARAMETER temperature 0.7
```

2. Create and run:

```bash
ollama create homedock -f Modelfile
ollama run homedock "What is HomeDock OS?"
```

### llama.cpp

```bash
./llama-cli -m HomeDock-OS-3B-Instruct.Q4_K_M.gguf \
  -p "<|im_start|>system\nYou are an expert assistant for HomeDock OS documentation.<|im_end|>\n<|im_start|>user\nWhat is Drop Zone?<|im_end|>\n<|im_start|>assistant\n" \
  -n 256
```

### LM Studio

1. Download the `.gguf` file
2. Import into LM Studio
3. Use the ChatML preset or configure manually

### Python (llama-cpp-python)

```python
from llama_cpp import Llama

llm = Llama(model_path="HomeDock-OS-3B-Instruct.Q4_K_M.gguf", n_ctx=2048)

response = llm.create_chat_completion(
    messages=[
        {"role": "system", "content": "You are an expert assistant for HomeDock OS documentation."},
        {"role": "user", "content": "How do I install Plex on HomeDock OS?"}
    ]
)
print(response["choices"][0]["message"]["content"])
```

## Example Conversations

**User:** What is HomeDock OS?

**Assistant:** HomeDock OS is a complete home cloud operating system that simplifies self-hosting. It provides a desktop-like interface in your browser with features like the App Store (200+ pre-configured applications), Drop Zone (encrypted file storage), Control Hub (container management), and automatic SSL configuration. It runs on Linux, Windows, macOS, and as a Docker container.

---

**User:** How do I configure SSL?

**Assistant:** HomeDock OS supports automatic SSL configuration. Navigate to Settings > SSL/HTTPS, enter your domain name, and HomeDock OS will automatically obtain and configure Let's Encrypt certificates. Applications installed through the App Store inherit SSL configuration automatically.

---

**User:** What is the difference between Plex and Jellyfin?

**Assistant:** Both are media server solutions available in the HomeDock OS App Store. Plex offers a polished interface with premium features like hardware transcoding and mobile sync (requires Plex Pass). Jellyfin is fully open-source and free, with no premium tier required. For most self-hosters, Jellyfin provides excellent functionality without subscription costs.

## Training Details

| Parameter | Value |
|-----------|-------|
| Base Model | Qwen/Qwen2.5-3B-Instruct |
| Method | LoRA (r=16, alpha=32) |
| Epochs | 3 |
| Learning Rate | 2e-4 |
| Sequence Length | 2048 |
| Training Examples | 788 |
| Total Words | ~99,000 |

### Training Data

The model was trained on the complete HomeDock OS documentation including:

- Official documentation (MDX)
- Application descriptions (200+ apps)
- Blog posts and tutorials
- Enterprise feature guides

## Limitations

- Knowledge is specific to HomeDock OS; may not generalize well to other platforms
- Training data cutoff corresponds to the documentation version at training time
- Best suited for English queries
- May occasionally generate plausible-sounding but incorrect information

## License

This model is released under the [Apache 2.0 License](https://www.apache.org/licenses/LICENSE-2.0), same as the base Qwen 2.5 model.

## Links

- [HomeDock OS Website](https://homedock.cloud)
- [HomeDock OS Documentation](https://docs.homedock.cloud)
- [Base Model: Qwen 2.5 3B Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct)

## Citation

```bibtex
@misc{homedock-os-3b-instruct,
  author = {Banshee Technologies S.L.},
  title = {HomeDock-OS-3B-Instruct: A Fine-tuned LLM for HomeDock OS Documentation},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/banshee/HomeDock-OS-3B-Instruct}
}
```

---

<p align="center">
  <b>Built with HomeDock OS</b><br>
  <a href="https://homedock.cloud">homedock.cloud</a>
</p>