---
language:
- en
- zh
library_name: transformers
tags:
- minicpm
- duplex
- text-generation
base_model: xinrongzhang2022/MiniCPM-duplex
---

# MiniCPM-duplex (safetensors)

Modern safetensors conversion of [xinrongzhang2022/MiniCPM-duplex](https://huggingface.co/xinrongzhang2022/MiniCPM-duplex).

**Weights are identical** — only the serialization format has changed from `pytorch_model.bin`
to `model.safetensors`, enabling memory-mapped loading and compatibility with current
versions of Transformers.

## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained(
    "enochlev/MiniCPM-duplex", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(
    "enochlev/MiniCPM-duplex",
    trust_remote_code=True,
    dtype=torch.float16,
    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))
```

## Original model

See [xinrongzhang2022/MiniCPM-duplex](https://huggingface.co/xinrongzhang2022/MiniCPM-duplex)
for the original weights, paper, and full documentation.