Instructions to use MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite with Transformers:

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

pipe = pipeline("text-generation", model="MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite")
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)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite")
model = AutoModelForMultimodalLM.from_pretrained("MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite", device_map="auto")
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=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite

SGLang

How to use MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite 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 "MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite" \
    --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": "MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite",
		"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 "MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite" \
        --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": "MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite with Docker Model Runner:
```
docker model run hf.co/MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite
```

SHUTEN-DOJI-CODE-27B-lite

SHUTEN-DOJI-CODE-27B-lite is a software engineering specialization of the SHUTEN-DŌJI model family developed by NULLXES.

The model is designed to improve software development workflows while preserving the strategic reasoning, planning, and analytical capabilities of the base SHUTEN-DOJI platform.

Capabilities

Code generation
Code review
Refactoring
Bug fixing
Unit test generation
Technical analysis
Engineering planning
Agent-oriented development workflows

Evaluation

Metric	Base Model	CODE Variant
Code Pass@1	0.0	0.4
Strategy Regression	—	0.0
Release Gate	—	PROMOTE

Training Objectives

This release focuses on improving engineering performance while maintaining existing strategic planning behavior.

Training objectives included:

Software engineering tasks
Code quality improvement
Execution-oriented outputs
Planning retention
Capability preservation

Architecture

SHUTEN-DOJI-CODE-27B-lite is built on top of the SHUTEN-DOJI platform and represents the first dedicated software engineering specialization within the SHUTEN ecosystem.

Current Status

Production candidate release.

Areas of future development:

Larger verified code datasets
Extended benchmark coverage
Multi-agent engineering workflows
Advanced repository reasoning
Long-horizon software planning

Notes

This release primarily targets software engineering and technical reasoning workloads.

Vision capabilities remain available through the underlying SHUTEN-DOJI platform and were not specifically adapted during this training cycle.

Downloads last month: 10

Safetensors

Model size

27B params

Tensor type

BF16

Model tree for MagistrTheOne/SHUTEN-DOJI-CODE-27B-lite

Base model

Qwen/Qwen3.6-27B

Finetuned

MagistrTheOne/SHUTEN-DOJI

Finetuned

(1)

this model