Instructions to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

llama-cpp-python

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf",
	filename="qwen2.5-coder-3b-vitest.Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M

Use Docker

docker model run hf.co/recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M

LM Studio
Jan

vLLM

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M

Ollama
How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with Ollama:
```
ollama run hf.co/recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M
```

Unsloth Studio new

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf to start chatting

Pi new

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf

Run Hermes

hermes

MLX LM

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Docker Model Runner
How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with Docker Model Runner:
```
docker model run hf.co/recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M
```

Lemonade

How to use recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf:Q4_K_M

Run and chat with the model

lemonade run user.qwen2.5-coder-3b-vitest.Q4_K_M.gguf-Q4_K_M

List all available models

lemonade list

Qwen2.5-Coder-3B-Vitest (GGUF)

A domain-specialized version of Qwen2.5-Coder-3B fine-tuned to generate Vitest + React Testing Library unit tests for React components.

The model focuses on behavior-faithful testing, avoiding hallucinated UI/state and preferring modern RTL best practices.

🔍 What this model does

Generates Vitest tests using React Testing Library.

Prefers:

getByRole, getByLabelText, getByPlaceholderText
userEvent for interactions
vi.fn() for callback assertions

Avoids:

Hallucinated UI or state
Testing behavior not present in the component
Legacy fireEvent

Tries to produce behavior-faithful, minimal tests.

🧠 Training details

Detail	Value
Base model	Qwen2.5-Coder-3B
Method	LoRA fine-tuning using MLX (Apple Silicon)
LoRA rank	16
Trainable parameters	~~6.6M (~~0.216% of total)
Dataset	~800 curated React component ↔ Vitest test pairs
Sequence length	2048
Training	~~3 epochs (~~450 steps)
Hardware	Apple Silicon (Mac mini M4 Pro, 64 GB RAM)

The dataset was curated and scored to:

Prefer robust RTL queries
Penalize async misuse, fireEvent, focused/skipped tests
Enforce correct callback testing for callback-only components
Encourage clean, production-style tests

📦 Files in this repo

qwen2.5-coder-3b-vitest.Q4_K_M.gguf — Quantized GGUF build (recommended for Ollama / llama.cpp)

🚀 Usage with Ollama

Create a Modelfile:

FROM qwen2.5-coder-3b-vitest.Q4_K_M.gguf
TEMPLATE """
You are an expert frontend test engineer.
{{ .Prompt }}
"""
PARAMETER temperature 0.2
PARAMETER top_p 0.9
PARAMETER num_ctx 4096

Then run:

ollama create vitest-coder -f Modelfile
ollama run vitest-coder "Write Vitest tests for this React component: ..."

⚠️ Limitations

May overuse async/await in simple sync cases
May generate redundant tests for trivial components
Best results when provided with clear component context
For large repos, using RAG (retrieval of similar tests/components) improves quality

📜 License

Same license as the base model: Qwen2.5-Coder-3B.

Downloads last month: 195

GGUF

Model size

3B params

Architecture

qwen2

Hardware compatibility

4-bit

Model tree for recursivecurse/qwen2.5-coder-3b-vitest.Q4_K_M.gguf

Base model

Qwen/Qwen2.5-3B

Finetuned

Qwen/Qwen2.5-Coder-3B

Adapter

(35)

this model