README.md

---
library_name: rkllm
pipeline_tag: text-generation
license: other
license_name: qwen-research
license_link: https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct/blob/main/LICENSE
base_model:
- Qwen/Qwen2.5-Coder-3B-Instruct
tags:
- text-generation-inference
- rkllm
- rk3588
- rockchip
- edge-ai
- qwen2
- code
- chat
---
# Qwen2.5-Coder-3B-Instruct — RKLLM build for RK3588 boards

### Built with Qwen

**Author:** @jamescallander  
**Source model:** [Qwen/Qwen2.5-Coder-3B-Instruct · Hugging Face](https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct)

**Target:** Rockchip RK3588 NPU via **RKNN-LLM Runtime**

> This repository hosts a **conversion** of `Qwen2.5-Coder-3B-Instruct` for use on Rockchip RK3588 single-board computers (Orange Pi 5 plus, Radxa Rock 5b+, Banana Pi M7, etc.). Conversion was performed using the [RKNN-LLM toolkit](https://github.com/airockchip/rknn-llm?utm_source=chatgpt.com)

#### Conversion details

- RKLLM-Toolkit version: v1.2.1
- NPU driver: v0.9.8
- Python: 3.12
- Quantization: `w8a8_g128`
- Output: single-file `.rkllm` artifact
- Modifications: quantization (w8a8_g128), export to .rkllm format for RK3588 SBCs.
- Tokenizer: not required at runtime (UI handles prompt I/O)

## ⚠️ Code generation disclaimer

🛑 **This model may produce incorrect, insecure, or non-optimal code.**

- It is intended for **research, educational, and prototyping purposes only**.

- Always **review, test, and validate** any generated code before using it in production.
- The model does not guarantee compliance with security best practices or coding standards.
- You are responsible for ensuring outputs meet your project’s requirements and legal obligations.
  
## Intended use

- On-device deployment of a **coding-focused instruction model** for software development assistance on SBCs.
- Qwen2.5-Coder-3B-Instruct is tuned for **code generation, explanation, and debugging tasks**, making it suitable for private edge inference.

## Limitations

- Requires 4GB free memory
- Quantized build (`w8a8_g128`) may show small quality differences vs. full-precision upstream.
- Tested on a Radxa Rock 5B+; other devices may require different drivers/toolkit versions.

## Quick start (RK3588)

### 1) Install runtime

The RKNN-LLM toolkit and instructions can be found on the specific development board's manufacturer website or from [airockchip's github page](https://github.com/airockchip).

Download and install the required packages as per the toolkit's instructions.

### 2) Simple Flask server deployment

The simplest way the deploy the `.rkllm` converted model is using an example script provided in the toolkit in this directory: `rknn-llm/examples/rkllm_server_demo`

```bash
python3 <TOOLKIT_PATH>/rknn-llm/examples/rkllm_server_demo/flask_server.py \
  --rkllm_model_path <MODEL_PATH>/Qwen2.5-Coder-3B-Instruct_w8a8_g128_rk3588.rkllm \
  --target_platform rk3588
```

### 3) Sending a request

A basic format for message request is:

```json
{
    "model":"Qwen2.5-Coder-3B",
    "messages":[{
        "role":"user",
        "content":"<YOUR_PROMPT_HERE>"}],
    "stream":false
}
```

Example request using `curl`:

```bash
curl -s -X POST <SERVER_IP_ADDRESS>:8080/rkllm_chat \
    -H 'Content-Type: application/json' \
    -d '{"model":"Qwen2.5-Coder-3B","messages":[{"role":"user","content":"Explain in one sentence  what a static method is."}],"stream":false}'
```

The response is formated in the following way:

```json
{
    "choices":[{
        "finish_reason":"stop",
        "index":0,
        "logprobs":null,
        "message":{
            "content":"<MODEL_REPLY_HERE">,
            "role":"assistant"}}],
        "created":null,
        "id":"rkllm_chat",
        "object":"rkllm_chat",
        "usage":{
            "completion_tokens":null,
            "prompt_tokens":null,
            "total_tokens":null}
}
```

Example response:

```json
{"choices":[{"finish_reason":"stop","index":0,"logprobs":null,"message":{"content":"A static method belongs to the class itself rather than any instance of the class and can be called without creating an object of the class.","role":"assistant"}}],"created":null,"id":"rkllm_chat","object":"rkllm_chat","usage":{"completion_tokens":null,"prompt_tokens":null,"total_tokens":null}}
```

### 4) UI compatibility

This server exposes an **OpenAI-compatible Chat Completions API**.

You can connect it to any OpenAI-compatible client or UI (for example: [Open WebUI](https://github.com/open-webui/open-webui?utm_source=chatgpt.com))
- Configure your client with the API base: `http://<SERVER_IP_ADDRESS>:8080` and use the endpoint: `/rkllm_chat`
- Make sure the `model` field matches the converted model’s name, for example:

```json
{
 "model": "Qwen2.5-Coder-3B-Instruct",
 "messages": [{"role":"user","content":"Hello!"}],
 "stream": false
}
```

# License

This conversion follows the license of the source model: [LICENSE · Qwen/Qwen2.5-Coder-3B-Instruct at main](https://huggingface.co/Qwen/Qwen2.5-Coder-3B-Instruct/blob/main/LICENSE)