Instructions to use OMNIRL/SPARK-SPLASH-4B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OMNIRL/SPARK-SPLASH-4B-GGUF with Transformers:

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

pipe = pipeline("text-generation", model="OMNIRL/SPARK-SPLASH-4B-GGUF")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("OMNIRL/SPARK-SPLASH-4B-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use OMNIRL/SPARK-SPLASH-4B-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0
# Run inference directly in the terminal:
llama cli -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0
# Run inference directly in the terminal:
llama cli -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

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 OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0
# Run inference directly in the terminal:
./llama-cli -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

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 OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

Use Docker

docker model run hf.co/OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

LM Studio
Jan

vLLM

How to use OMNIRL/SPARK-SPLASH-4B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OMNIRL/SPARK-SPLASH-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": "OMNIRL/SPARK-SPLASH-4B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

SGLang

How to use OMNIRL/SPARK-SPLASH-4B-GGUF 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 "OMNIRL/SPARK-SPLASH-4B-GGUF" \
    --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": "OMNIRL/SPARK-SPLASH-4B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "OMNIRL/SPARK-SPLASH-4B-GGUF" \
        --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": "OMNIRL/SPARK-SPLASH-4B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use OMNIRL/SPARK-SPLASH-4B-GGUF with Ollama:
```
ollama run hf.co/OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0
```
Unsloth Desktop

How to use OMNIRL/SPARK-SPLASH-4B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use OMNIRL/SPARK-SPLASH-4B-GGUF with Docker Model Runner:
```
docker model run hf.co/OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0
```

Lemonade

How to use OMNIRL/SPARK-SPLASH-4B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

Run and chat with the model

lemonade run user.SPARK-SPLASH-4B-GGUF-Q8_0

List all available models

lemonade list

Hermes Agent

How to use OMNIRL/SPARK-SPLASH-4B-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

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 OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use OMNIRL/SPARK-SPLASH-4B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "OMNIRL/SPARK-SPLASH-4B-GGUF:Q8_0" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Spark Series – Splash & Pro Aquatic

Advanced Agentic Language Models for Reasoning, Coding, and Intelligent Workflows.

Spark Series is a family of efficient, agentic LLMs built for long-context reasoning, tool use, coding, mathematics, and autonomous workflows.

Spark Series Splash: Fast & efficient model – 4B parameters
Spark Series Pro Aquatic: Flagship model – most powerful, multimodal + agentic — 🚧 Coming Soon

✨ Models Overview

1. Spark Series Splash

Advanced agentic language model designed for long-context reasoning, tool usage, coding, mathematics, problem-solving, and intelligent workflows.

Optimized for efficiency without sacrificing performance.

Specs:

Feature	Detail
Parameters	4B
Context	Long-context + memory retention
Capabilities	Reasoning, Coding, Math, Tool Calling

Key Features:

Agentic reasoning and autonomous task execution
Long-context understanding and memory retention
Advanced coding and software development
Strong mathematical and logical reasoning
Tool calling and workflow automation
Optimized for speed, efficiency, and intelligence

2. Spark Series Pro Aquatic — 🚧 Coming Soon

🚧 Coming Soon – Pro Aquatic is not released yet. Stay tuned!

The most powerful model in the Spark Series family. Cutting-edge multimodal + agentic system capable of understanding text, images, documents, and complex workflows.

Designed for professional users, developers, researchers, and enterprises.

Key Features:

Most powerful model in the lineup
Native multimodal understanding (text, images, documents)
Advanced agentic reasoning and planning
Long-context processing
High-performance coding and software engineering
Professional-grade math and scientific reasoning
Autonomous tool calling and workflow execution
Enhanced instruction following
Enterprise-level reliability and scalability

🚀 Intended Use

Intended for:

Agentic assistants and copilots
Code generation, debugging, refactoring
Math / STEM problem solving
Document Q&A and summarization (Pro Aquatic)
Tool calling, function calling, workflow automation
Research, prototyping, enterprise apps

Out-of-scope:

No guarantee of factual accuracy – always verify critical outputs
Not for high-risk decisions without human oversight

💻 How to Use

With `transformers`:

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "OMNIRL/SPARK-SPLASH-4B"  # or spark-series-pro-aquatic

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    device_map="auto",
    trust_remote_code=True
)

messages = [
    {"role": "system", "content": "You are Spark, a helpful agentic assistant."},
    {"role": "user", "content": "Write a Python function to sort a list and explain it in one line."}
]

inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True).to(model.device)
outputs = model.generate(inputs, max_new_tokens=512, temperature=0.7, do_sample=True)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

API Access

API access: $11 per API call token

Want to try Spark Series Pro Aquatic for free? Create an account on the Spark Series platform and get 10 days of free trial access.

⚠️ Limitations & Bias

May hallucinate or produce incorrect code/math – test outputs.
Multimodal (Pro Aquatic) performance depends on input quality.
English-optimized, multilingual ability may vary.
Generation is fast but sampling settings affect quality.

📄 License

Apache-2.0 – See LICENSE for details.

If you use this work, please cite:

@misc{sparkseries2026,
  title={Spark Series: Splash and Pro Aquatic, Agentic Language Models},
  author={Spark Series Team},
  year={2026}
}

🔗 Links & Contact

Hugging Face: OMNIRL/SPARK-SPLASH-4B
Issues / Requests: Open a Discussion or Issue in this repo
Trial: Spark Series Platform – 10 days free for Pro Aquatic

Downloads last month: 185

GGUF

Model size

4B params

Architecture

qwen35

Hardware compatibility

8-bit

Spark Series – Splash & Pro Aquatic

✨ Models Overview

1. Spark Series Splash

2. Spark Series Pro Aquatic — 🚧 Coming Soon

🚀 Intended Use

💻 How to Use

With transformers:

API Access

⚠️ Limitations & Bias

📄 License

🔗 Links & Contact

With `transformers`: