Instructions to use OceanLabs/Ocean-1-4B-DSpark with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OceanLabs/Ocean-1-4B-DSpark with Transformers:

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

pipe = pipeline("text-generation", model="OceanLabs/Ocean-1-4B-DSpark", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("OceanLabs/Ocean-1-4B-DSpark", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OceanLabs/Ocean-1-4B-DSpark with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OceanLabs/Ocean-1-4B-DSpark"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OceanLabs/Ocean-1-4B-DSpark",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/OceanLabs/Ocean-1-4B-DSpark

SGLang

How to use OceanLabs/Ocean-1-4B-DSpark 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 "OceanLabs/Ocean-1-4B-DSpark" \
    --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": "OceanLabs/Ocean-1-4B-DSpark",
		"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 "OceanLabs/Ocean-1-4B-DSpark" \
        --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": "OceanLabs/Ocean-1-4B-DSpark",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use OceanLabs/Ocean-1-4B-DSpark with Docker Model Runner:
```
docker model run hf.co/OceanLabs/Ocean-1-4B-DSpark
```

D-Spark-4B

D-Spark-4B is a compact, general-purpose language model optimized for efficient inference on small devices while preserving strong performance across conversation, writing, translation, reasoning, coding, tool use, and agentic workflows. It natively supports context windows of up to 1M tokens and more than 200 languages.

Technical Highlights

Efficient Hybrid Architecture: Combines sliding-window attention (majority of layers) with sparse full-attention layers. This design significantly reduces KV-cache memory and compute on resource-constrained hardware while retaining long-context capability.
Optimized for Edge & On-Device: Reduced intermediate size and a higher ratio of sliding-window layers improve TTFT, throughput, and memory footprint compared with similar 4B-class models.
Strong Coding & Agent Capabilities: Compatible with popular agent frameworks. Delivers competitive results on everyday coding, agentic workflows, reasoning, and instruction-following tasks.
Broad Compatibility: Supports NVIDIA, Apple Silicon (MLX), CPU, and other platforms. Works with vLLM, SGLang, llama.cpp, Ollama, LM Studio, and can be fine-tuned with Llama-Factory.

Model Overview

Parameter	Value
Parameters	~3.9–4.0 B
Hidden size	2560
Intermediate size	9216
Layers	36
Attention heads	16 (KV: 4)
Head dim	256
Sliding window	512
Max context	1 048 576
Vocab size	131 072
Precision	bfloat16

Layer pattern (higher sliding ratio for efficiency): 4 × sliding → 1 × full (repeated), ending with sliding.

Recommended Sampling

{
  "temperature": 1.0,
  "top_p": 0.95,
  "top_k": -1,
  "repetition_penalty": 1.0
}

Thinking mode is enabled by default via the chat template. Disable with "chat_template_kwargs": {"enable_thinking": false}.

Quickstart (Transformers)

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "path/to/D-Spark-4B"
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True,
)

messages = [{"role": "user", "content": "What is the capital of Anhui Province?"}]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt", add_generation_prompt=True).to(model.device)
outputs = model.generate(inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

License

Apache 2.0

Citation

@misc{dspark,
    title  = {D-Spark-4B: Efficient On-Device Language Model},
    author = {D-Spark Team},
    year   = {2026}
}

Downloads last month: 425

Safetensors

Model size

4B params

Tensor type

BF16

Collection including OceanLabs/Ocean-1-4B-DSpark

Ocean-1

Collection

3 items • Updated about 4 hours ago • 1