Instructions to use enochlev/MiniCPM-duplex with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use enochlev/MiniCPM-duplex with Transformers:

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

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

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("enochlev/MiniCPM-duplex", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use enochlev/MiniCPM-duplex with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/enochlev/MiniCPM-duplex

SGLang

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

Docker Model Runner
How to use enochlev/MiniCPM-duplex with Docker Model Runner:
```
docker model run hf.co/enochlev/MiniCPM-duplex
```

enochlev commited on 1 day ago

Commit

c19cdc1

verified ·

1 Parent(s): 30747ee

Clean up README: remove patch notes, fix prompt display

Browse files

Files changed (1) hide show

README.md +1 -8

README.md CHANGED Viewed

@@ -18,13 +18,6 @@ Modern safetensors conversion of [xinrongzhang2022/MiniCPM-duplex](https://huggi
 to `model.safetensors`, enabling memory-mapped loading and compatibility with current
 versions of Transformers.
-## Compatibility patch
-The original `modeling_minicpm.py` was written against an older Transformers Cache API.
-The copy here is patched for `transformers >= 4.38`: removed `DynamicCache` methods
-`seen_tokens`, `get_max_length()`, and `get_usable_length()` are replaced with
-`get_seq_length()`.
 ## Usage
 ```python
@@ -40,7 +33,7 @@ model = AutoModelForCausalLM.from_pretrained(
     device_map="auto",
 )
-prompt = "<\u7528\u6237>Hello, what can you do?<AI>"
 inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
 out = model.generate(**inputs, max_new_tokens=256)
 print(tokenizer.decode(out[0], skip_special_tokens=True))

 to `model.safetensors`, enabling memory-mapped loading and compatibility with current
 versions of Transformers.
 ## Usage
 ```python
     device_map="auto",
 )
+prompt = "<用户>Hello, what can you do?<AI>"
 inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
 out = model.generate(**inputs, max_new_tokens=256)
 print(tokenizer.decode(out[0], skip_special_tokens=True))