Instructions to use tiny-random/minicpm4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tiny-random/minicpm4 with Transformers:

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

pipe = pipeline("text-generation", model="tiny-random/minicpm4", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("tiny-random/minicpm4", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use tiny-random/minicpm4 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tiny-random/minicpm4"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tiny-random/minicpm4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tiny-random/minicpm4

SGLang

How to use tiny-random/minicpm4 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 "tiny-random/minicpm4" \
    --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": "tiny-random/minicpm4",
		"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 "tiny-random/minicpm4" \
        --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": "tiny-random/minicpm4",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tiny-random/minicpm4 with Docker Model Runner:
```
docker model run hf.co/tiny-random/minicpm4
```

minicpm4 / config.json

yujiepan

Upload folder using huggingface_hub

f1885fe verified 11 months ago

raw

history blame contribute delete

2.18 kB

	{
	"architectures": [
	"MiniCPMForCausalLM"
	],
	"attention_bias": false,
	"attention_dropout": 0.0,
	"auto_map": {
	"AutoConfig": "openbmb/MiniCPM4-8B--configuration_minicpm.MiniCPMConfig",
	"AutoModel": "openbmb/MiniCPM4-8B--modeling_minicpm.MiniCPMModel",
	"AutoModelForCausalLM": "openbmb/MiniCPM4-8B--modeling_minicpm.MiniCPMForCausalLM",
	"AutoModelForSeq2SeqLM": "openbmb/MiniCPM4-8B--modeling_minicpm.MiniCPMForCausalLM",
	"AutoModelForSequenceClassification": "openbmb/MiniCPM4-8B--modeling_minicpm.MiniCPMForSequenceClassification"
	},
	"bos_token_id": 1,
	"dim_model_base": 32,
	"eos_token_id": [
	2,
	73440
	],
	"hidden_act": "silu",
	"hidden_size": 64,
	"initializer_range": 0.1,
	"intermediate_size": 128,
	"max_position_embeddings": 32768,
	"model_type": "minicpm",
	"mup_denominator": 32,
	"num_attention_heads": 2,
	"num_hidden_layers": 2,
	"num_key_value_heads": 1,
	"pad_token_id": 2,
	"pretraining_tp": 1,
	"rms_norm_eps": 1e-06,
	"rope_scaling": {
	"long_factor": [
	0.9977997200264581,
	1.014658295992452,
	1.0349680404997148,
	1.059429246056193,
	1.0888815016813513,
	1.1243301355211495,
	1.166977103606075,
	1.2182568066927284,
	1.2798772354275727,
	1.3538666751582975,
	1.4426259039919596,
	1.5489853358570191,
	1.6762658237220625,
	1.8283407612492941,
	2.0096956085876183,
	2.225478927469756
	],
	"original_max_position_embeddings": 32768,
	"rope_type": "longrope",
	"short_factor": [
	0.9977997200264581,
	1.014658295992452,
	1.0349680404997148,
	1.059429246056193,
	1.0888815016813513,
	1.1243301355211495,
	1.166977103606075,
	1.2182568066927284,
	1.2798772354275727,
	1.3538666751582975,
	1.4426259039919596,
	1.5489853358570191,
	1.6762658237220625,
	1.8283407612492941,
	2.0096956085876183,
	2.225478927469756
	]
	},
	"rope_theta": 10000.0,
	"scale_depth": 1.4,
	"scale_emb": 12,
	"sparse_config": null,
	"torch_dtype": "bfloat16",
	"transformers_version": "4.51.3",
	"use_cache": true,
	"vocab_size": 73448
	}