Instructions to use WithinUsAI/WithIn-Us-Coder-4B.gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use WithinUsAI/WithIn-Us-Coder-4B.gguf with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="WithinUsAI/WithIn-Us-Coder-4B.gguf",
	filename="WithIn-Us-Coder-4B.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 WithinUsAI/WithIn-Us-Coder-4B.gguf with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf WithinUsAI/WithIn-Us-Coder-4B.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 WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf WithinUsAI/WithIn-Us-Coder-4B.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 WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf WithinUsAI/WithIn-Us-Coder-4B.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 WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M

Use Docker

docker model run hf.co/WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M

LM Studio
Jan

vLLM

How to use WithinUsAI/WithIn-Us-Coder-4B.gguf with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "WithinUsAI/WithIn-Us-Coder-4B.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": "WithinUsAI/WithIn-Us-Coder-4B.gguf",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M

Ollama
How to use WithinUsAI/WithIn-Us-Coder-4B.gguf with Ollama:
```
ollama run hf.co/WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M
```

Unsloth Studio new

How to use WithinUsAI/WithIn-Us-Coder-4B.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 WithinUsAI/WithIn-Us-Coder-4B.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 WithinUsAI/WithIn-Us-Coder-4B.gguf to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for WithinUsAI/WithIn-Us-Coder-4B.gguf to start chatting

Pi new

How to use WithinUsAI/WithIn-Us-Coder-4B.gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use WithinUsAI/WithIn-Us-Coder-4B.gguf with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M

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 WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use WithinUsAI/WithIn-Us-Coder-4B.gguf with Docker Model Runner:
```
docker model run hf.co/WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M
```

Lemonade

How to use WithinUsAI/WithIn-Us-Coder-4B.gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull WithinUsAI/WithIn-Us-Coder-4B.gguf:Q4_K_M

Run and chat with the model

lemonade run user.WithIn-Us-Coder-4B.gguf-Q4_K_M

List all available models

lemonade list

WithIn-Us-Coder-4B.gguf

WithIn-Us-Coder-4B.gguf is a GGUF release from WithIn Us AI, built for local inference and coding-focused assistant use cases. It is based on Qwen/Qwen3.5-4B and distributed in quantized GGUF formats for efficient deployment in llama.cpp-compatible runtimes.

Model Summary

This model is intended as a coding-oriented conversational assistant with emphasis on:

code generation
code reasoning
implementation planning
debugging assistance
instruction following
general assistant-style chat for development workflows

This repository currently provides the following GGUF variants:

WithIn-Us-Coder-4B.Q4_K_M.gguf
WithIn-Us-Coder-4B.Q5_K_M.gguf

Creator

WithIn Us AI is the creator of this model release, including the model packaging, fine-tuning / merging concept, process, naming, and GGUF distribution.

Base Model

This model is based on:

Qwen/Qwen3.5-4B

Credit and appreciation go to the original creators of the base LLM architecture and weights.

Training Data

The current repository metadata lists the following datasets as part of the model’s training / fine-tuning lineage:

WithinUsAI/Python_GOD_Coder_50k
reedmayhew/gemini-3.1-pro-2048-reasoning-1100x
m-a-p/Code-Feedback
crownelius/Opus-4.6-Reasoning-2100x-formatted
crownelius/Opus4.6-No-Reasoning-260x
crownelius/Creative_Writing_Multiturn_Enhanced
HuggingFaceH4/llava-instruct-mix-vsft
Roman1111111/gemini-3-pro-10000x-hard-high-reasoning

Attribution note:
WithIn Us AI does not claim ownership over third-party base models or third-party datasets. Full credit, thanks, and attribution belong to the original model and dataset creators.

Intended Use

This model is intended for:

local coding assistants
offline development help
code explanation
bug-fixing support
prompt-based code generation
experimentation in llama.cpp and GGUF-compatible environments

Suggested Use Cases

generating Python, JavaScript, C++, and other programming language snippets
explaining code blocks
rewriting or improving functions
brainstorming implementation strategies
creating scaffolding and prototypes
assisting with debugging and refactoring

Out-of-Scope Use

This model is not guaranteed to be reliable for:

high-stakes legal advice
medical advice
financial decision-making
autonomous execution without review
security-critical production decisions without human verification

Users should always validate generated code before deployment.

Quantization Formats

This repository currently includes:

Q4_K_M for smaller memory footprint and faster local inference
Q5_K_M for improved quality while remaining efficient

Choose the quant level based on your hardware budget and quality needs.

Prompting Notes

As a coding-focused conversational model, best results usually come from prompts that are:

specific
structured
explicit about language, framework, or goal
clear about desired output format

Example prompt style:

Write a Python function that parses a CSV file, removes duplicate rows by email, and saves the cleaned result. Include error handling and comments.

Limitations

Like other language models, this model may:

hallucinate APIs or library behavior
generate insecure or inefficient code
make reasoning mistakes
produce outdated patterns
require prompt iteration for best results

Human review is strongly recommended, especially for production code.

License

This repository uses a custom WithIn Us AI license approach.

The base model may be subject to its original upstream license and terms.
Third-party datasets remain the property of their respective creators / licensors.
WithIn Us AI claims authorship of the fine-tuning / merging concept, process, packaging, naming, and release structure for this model distribution.
This repository does not claim ownership over third-party datasets or the underlying upstream base model.

You can include a LICENSE file in this repository with the exact custom terms you want enforced.

Acknowledgments

Special thanks to:

Qwen for the base model
all third-party dataset creators listed above
the open-source GGUF / llama.cpp ecosystem
the broader Hugging Face community

Files

Current repository files include:

WithIn-Us-Coder-4B.Q4_K_M.gguf
WithIn-Us-Coder-4B.Q5_K_M.gguf

Disclaimer

This model may generate incorrect, biased, insecure, or incomplete outputs.
Use responsibly, validate important results, and review all generated code before real-world use.