Instructions to use axteam/MiniCPM5-2B-AWQ-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use axteam/MiniCPM5-2B-AWQ-4bit with Transformers:

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

pipe = pipeline("text-generation", model="axteam/MiniCPM5-2B-AWQ-4bit")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("axteam/MiniCPM5-2B-AWQ-4bit")
model = AutoModelForCausalLM.from_pretrained("axteam/MiniCPM5-2B-AWQ-4bit", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use axteam/MiniCPM5-2B-AWQ-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "axteam/MiniCPM5-2B-AWQ-4bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "axteam/MiniCPM5-2B-AWQ-4bit",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/axteam/MiniCPM5-2B-AWQ-4bit

SGLang

How to use axteam/MiniCPM5-2B-AWQ-4bit 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 "axteam/MiniCPM5-2B-AWQ-4bit" \
    --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": "axteam/MiniCPM5-2B-AWQ-4bit",
		"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 "axteam/MiniCPM5-2B-AWQ-4bit" \
        --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": "axteam/MiniCPM5-2B-AWQ-4bit",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use axteam/MiniCPM5-2B-AWQ-4bit with Docker Model Runner:
```
docker model run hf.co/axteam/MiniCPM5-2B-AWQ-4bit
```

MiniCPM5-2B · AWQ INT4 (W4A16)

4-bit quantization of openbmb/MiniCPM5-2B for vLLM serving. Quantized by axteam.


Base model	`openbmb/MiniCPM5-2B`
Method	AWQ INT4 (W4A16)
Format	compressed-tensors `pack-quantized`
Weights / activations	W4A16 — INT4 asym weights, group 32, BF16 activations
Architecture	Dense (Llama-architecture) · text-only
Size on disk	~2.1 GB
Calibration	256 samples
License	Apache-2.0

Quantization details

Scheme — W4A16 — INT4 asym weights, group 32, BF16 activations, stored as compressed-tensors pack-quantized.
Kept in BF16 — token embeddings, the LM head, and all RMSNorms (a small fraction of parameters).
Post-training quantization (no retraining). Tokenizer / chat_template identical to the original.

Evaluation (measured)

Fidelity to the BF16 original on an internal mixed EN/KO set (100 prompts, BF16 greedy answers scored under each build).

Metric	This model	BF16 original
Perplexity (lower = closer)	1.8260	1.5988 (+14.2%)
Top-1 next-token agreement	84.9%	98.9%

⚠️ Note: noticeable 4-bit degradation here; prefer the 8-bit (FP8 / AWQ INT8) builds.

Usage (vLLM)

vllm serve axteam/MiniCPM5-2B-AWQ-4bit

License

Quantized derivative of openbmb/MiniCPM5-2B (Apache-2.0). Original weights & tokenizer belong to OpenBMB.

한국어 요약

openbmb/MiniCPM5-2B 를 AWQ INT4 (W4A16) 로 양자화한 서빙용(4-bit) 모델입니다. 작성: axteam.

방식 — INT4 비대칭 32묶음, 활성화 BF16(W4A16). 형식 compressed-tensors pack-quantized.
BF16 유지 — 토큰 임베딩·LM 헤드·모든 norm.
PTQ(재학습 없음), 토크나이저·chat_template 원본 동일.
품질(실측) — BF16 대비 perplexity 1.8260(원본 1.5988, +14.2%), top-1 84.9%(원본 98.9%). 내부 영/한 100문항.
⚠️ 4비트 열화가 큼 — 8비트(FP8/AWQ INT8) 권장.
사용 — vllm serve axteam/MiniCPM5-2B-AWQ-4bit. 라이선스 Apache-2.0(원본 OpenBMB).