Instructions to use openbmb/MiniCPM4-8B-marlin-cpmcu with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use openbmb/MiniCPM4-8B-marlin-cpmcu with Transformers:

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

pipe = pipeline("text-generation", model="openbmb/MiniCPM4-8B-marlin-cpmcu", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("openbmb/MiniCPM4-8B-marlin-cpmcu", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use openbmb/MiniCPM4-8B-marlin-cpmcu with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "openbmb/MiniCPM4-8B-marlin-cpmcu"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openbmb/MiniCPM4-8B-marlin-cpmcu",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/openbmb/MiniCPM4-8B-marlin-cpmcu

SGLang

How to use openbmb/MiniCPM4-8B-marlin-cpmcu 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 "openbmb/MiniCPM4-8B-marlin-cpmcu" \
    --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": "openbmb/MiniCPM4-8B-marlin-cpmcu",
		"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 "openbmb/MiniCPM4-8B-marlin-cpmcu" \
        --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": "openbmb/MiniCPM4-8B-marlin-cpmcu",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use openbmb/MiniCPM4-8B-marlin-cpmcu with Docker Model Runner:
```
docker model run hf.co/openbmb/MiniCPM4-8B-marlin-cpmcu
```

Improve model card: Add paper/project links and `code-generation`, `datasets`, `metrics` tags

by nielsr HF Staff - opened Sep 5, 2025

base: refs/heads/main

←

from: refs/pr/1

Discussion Files changed

+348

-123

nielsr

Sep 5, 2025

This PR enhances the model card for MiniCPM4-8B by:

Adding direct links to the official Hugging Face paper page (MiniCPM4: Ultra-Efficient LLMs on End Devices) and the Hugging Face collection project page (https://huggingface.co/collections/openbmb/minicpm4-6841ab29d180257e940baa9b) in the header section. Existing links pointing to the GitHub technical report PDF in "What's New" and "Citation" are updated to the HF paper link. The BibTeX entry is also updated with the arXiv journal.
Introducing the code-generation tag to the metadata, as supported by the model's performance on code benchmarks (HumanEval+, MBPP+, LiveCodeBench) and the "MCP-Code-Executor" application mentioned in the GitHub README.
Adding the datasets tag for openbmb/Ultra-FineWeb and the metrics tag for accuracy, as both are explicitly mentioned and demonstrated in the model's documentation.
Restructuring the "Usage" section to align with the original GitHub README's presentation of InfLLM v2 and rope_scaling details under the Hugging Face inference, rather than the CPM.cu section. This ensures consistency and clarity for users.

These improvements will make the model more discoverable, accurately categorized, and provide a clearer usage guide on the Hugging Face Hub.

Improve model card: Add paper/project links and `code-generation`, `datasets`, `metrics` tagsb54b9250

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

Tap or paste here to upload images

Ready to merge

This branch is ready to get merged automatically.

· Sign up or log in to comment