Instructions to use TitleOS/Spark-270M-FP16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TitleOS/Spark-270M-FP16 with Transformers:

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

pipe = pipeline("text-generation", model="TitleOS/Spark-270M-FP16")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("TitleOS/Spark-270M-FP16")
model = AutoModelForCausalLM.from_pretrained("TitleOS/Spark-270M-FP16")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.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(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use TitleOS/Spark-270M-FP16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TitleOS/Spark-270M-FP16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TitleOS/Spark-270M-FP16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/TitleOS/Spark-270M-FP16

SGLang

How to use TitleOS/Spark-270M-FP16 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 "TitleOS/Spark-270M-FP16" \
    --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": "TitleOS/Spark-270M-FP16",
		"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 "TitleOS/Spark-270M-FP16" \
        --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": "TitleOS/Spark-270M-FP16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use TitleOS/Spark-270M-FP16 with Docker Model Runner:
```
docker model run hf.co/TitleOS/Spark-270M-FP16
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Spark-270M

Spark-270M is a highly compact, utility-focused language model with 270 million parameters. It is a fine-tune of Google's Gemma 3 270M, designed to punch significantly above its weight class by leveraging high-quality synthetic data distillation.

The model functions as a "dense information engine"—specializing in generating concise title summaries, search engine queries, and logical follow-up questioning—while retaining the creative conversational flair inherited from its teacher model's lineage.

⚡ Model Details

Model Name: Spark-270M
Base Architecture: Google Gemma 3 270M
Parameters: 270M active parameters
Context Window: 32k tokens
Teacher Model: Lightning-1.7B (Custom model fine-tuned on Hermes 3)
Training Type: Synthetic "Textbook" Distillation (SFT)

📚 Training Methodology: "Textbooks Are All You Need"

Spark-270M was trained using a distinct data pipeline inspired by the Textbooks Are All You Need (Microsoft Phi) research paper.

Instead of training on raw web scrapes, Spark-270M was fine-tuned exclusively on a series of synthetic textbooks generated by a larger parent model, Lightning-1.7B.

The Teacher: Lightning-1.7B

The data generator, Lightning-1.7B, was itself fine-tuned on the Hermes 3 dataset. This lineage allows Spark-270M to inherit specific behavioral traits from Hermes 3—namely creativity, steerability, and a refusal to be "boring"—despite being distilled into a rigid textbook format.

The synthetic data focused on:

High-density reasoning chains: Explaining complex topics in compressed formats.
Utility Tasks: Converting conversational fluff into actionable queries.
Socratic Dialogue: Modeling inquisitive follow-up questioning.

🛠️ Intended Use & Capabilities

Spark-270M is designed to be a lightweight Utility Model. It is ideal for edge devices, rapid prototyping, or functioning as a specific "node" in a larger agentic system (e.g., the summarizer node or the query-generator node).

Primary Capabilities

Dense Title Summarization: Converting long conversation threads into information-dense, short titles or abstracts.
Search Query Generation: Formulating precise, keyword-rich search queries based on vague user input.
Proactive Questioning: Generating relevant follow-up questions to clarify user intent or deepen a topic.

💻 Example Usage

from transformers import AutoTokenizer, AutoModelForCausalLM

model_id = "TitleOS/Spark-270M"

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

# Example: Generating a search query from a user problem
input_text = """
User: I need to fix my sink, it's leaking from the bottom pipe where the U-shape thing is.
Task: Generate 3 search engine queries for this problem.
Response:
"""

input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")

outputs = model.generate(**input_ids, max_new_tokens=128)
print(tokenizer.d ecode(outputs[0]))