Instructions to use CodeAid/CouplingSmells-Detection-model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use CodeAid/CouplingSmells-Detection-model with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="CodeAid/CouplingSmells-Detection-model")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("CodeAid/CouplingSmells-Detection-model")
model = AutoModelForCausalLM.from_pretrained("CodeAid/CouplingSmells-Detection-model")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use CodeAid/CouplingSmells-Detection-model with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CodeAid/CouplingSmells-Detection-model"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CodeAid/CouplingSmells-Detection-model",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/CodeAid/CouplingSmells-Detection-model

SGLang

How to use CodeAid/CouplingSmells-Detection-model with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "CodeAid/CouplingSmells-Detection-model" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CodeAid/CouplingSmells-Detection-model",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "CodeAid/CouplingSmells-Detection-model" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CodeAid/CouplingSmells-Detection-model",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use CodeAid/CouplingSmells-Detection-model with Docker Model Runner:
```
docker model run hf.co/CodeAid/CouplingSmells-Detection-model
```

CouplingSmells-Detection-model

File size: 1,173 Bytes

cbdb119

---
license: apache-2.0
model_type: qwen
datasets:
- CodeAid/CouplingDetectionData
language:
- en
base_model:
- Qwen/Qwen2.5-14B-Instruct
pipeline_tag: text-generation
library_name: transformers
tags:
- llms
- code
- Java
- code-smells
- transformers
---

# CodeAid Coupling Smells Detection Model (Qwen2.5-14B-Instruct Fine-Tuned)

This model is a fine-tuned version of [Qwen2.5-14B-Instruct](https://huggingface.co/Qwen/Qwen2.5-14B-Instruct), specialized for detecting **coupling smells** in Java code. It was developed as part of the CodeAid project to assist developers in identifying code quality issues directly in their IDE.

## 🧠 Model Purpose

The model identifies **coupling-related code smells** such as:

- **Feature Envy**
- **Inappropriate Intimacy**
- **Message Chains**
- **Excessive Dependencies**

It analyzes Java classes and their dependencies to detect architectural or design issues that increase coupling and reduce maintainability.

## 🔧 Technical Details

- **Base Model**: Qwen2.5-14B-Instruct
- **Fine-Tuning Method**: QLoRA with LoRA adapters merged
- **Format**: `safetensors` (merged)
- **Task Type**: Text generation (instruction-based)