Instructions to use MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured with Transformers:

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

pipe = pipeline("text-generation", model="MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured")
model = AutoModelForCausalLM.from_pretrained("MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured")
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 MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured

SGLang

How to use MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured 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 "MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured" \
    --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": "MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured",
		"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 "MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured" \
        --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": "MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured with Docker Model Runner:
```
docker model run hf.co/MasterControlAIML/DeepSeek-R1-Strategy-Qwen-2.5-1.5b-Unstructured-To-Structured
```

Prompt example

by Safreliy - opened Feb 3, 2025

Discussion

Safreliy

Feb 3, 2025

Can you please provide an example of prompt with unstructured data, model schema and instructions for generating structured output?

bhaviktheslider

MasterControl org Feb 3, 2025

Hey @Safreliy You can look at the dataset: https://huggingface.co/datasets/MasterControlAIML/JSON-Unstructured-Structured , you have to give model some text, a schema to follow in the prompt and it will create a structured/json output for you.

Also we did further training and you can find update model here: https://huggingface.co/MasterControlAIML/DeepSeek-R1-Qwen-2.5-1.5b-Latest-Unstructured-To-Structured

Let me know if you have any more questions!

Safreliy

Feb 4, 2025

@bhaviktheslider Thank you! Is new version better in terms of val loss, or some benchmark metrics? Your model looks nice, really interested in it

bhaviktheslider

MasterControl org Feb 4, 2025

•

edited Feb 4, 2025

@Safreliy It was trained on more steps with improved prompt and reward metrics for format and answering in json. Convergence for reward gain can be seen in this one in a good way, and the way it controlled its completion length to a good amount by the end of 450 steps. You can find these metrics in the tensorboard section of the model.

Tensorboard section of new model: https://huggingface.co/MasterControlAIML/DeepSeek-R1-Qwen-2.5-1.5b-Latest-Unstructured-To-Structured/tensorboard

Thank you!

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment