Instructions to use wangkevin02/USP with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use wangkevin02/USP with Transformers:

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

pipe = pipeline("text-generation", model="wangkevin02/USP")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("wangkevin02/USP")
model = AutoModelForCausalLM.from_pretrained("wangkevin02/USP")
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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use wangkevin02/USP with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "wangkevin02/USP"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "wangkevin02/USP",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/wangkevin02/USP

SGLang

How to use wangkevin02/USP 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 "wangkevin02/USP" \
    --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": "wangkevin02/USP",
		"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 "wangkevin02/USP" \
        --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": "wangkevin02/USP",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use wangkevin02/USP with Docker Model Runner:
```
docker model run hf.co/wangkevin02/USP
```

wangkevin02 commited on Feb 27, 2025

Commit

aebe360

verified ·

1 Parent(s): a091ea3

Upload README.md with huggingface_hub

Browse files

Files changed (1) hide show

README.md +10 -2

README.md CHANGED Viewed

@@ -13,7 +13,7 @@ base_model:
 > **GitHub repository** for exploring the source code and additional resources: https://github.com/wangkevin02/USP
-Our User Simulator with Implicit Profiles (USP) replicates human-like conversational behavior in interactions with large language models (LLMs). By emulating diverse user dynamics based on predefined profiles, it reconstructs realistic user-LLM dialogues, leveraging the LLaMA-3-base-8B architecture with **Conditional Supervised Fine-Tuning (SFT)** and **Reinforcement Learning with Cycle Consistency (RLCC)**.  For a detailed methodology and insights, refer to [Our Paper](#).
 > *Note*: Our model is subject to the following constraints:
 >
@@ -97,6 +97,14 @@ print(f"Generated response: {responses}")
 If you find this model useful, please cite:
 ```plaintext
-[Authors], "[Paper Title]," [Venue], [Year], [URL or DOI].
 ```

 > **GitHub repository** for exploring the source code and additional resources: https://github.com/wangkevin02/USP
+Our User Simulator with Implicit Profiles (USP) replicates human-like conversational behavior in interactions with large language models (LLMs). By emulating diverse user dynamics based on predefined profiles, it reconstructs realistic user-LLM dialogues, leveraging the LLaMA-3-base-8B architecture with **Conditional Supervised Fine-Tuning (SFT)** and **Reinforcement Learning with Cycle Consistency (RLCC)**.  For a detailed methodology and insights, refer to [Our Paper](https://arxiv.org/pdf/2502.18968).
 > *Note*: Our model is subject to the following constraints:
 >
 If you find this model useful, please cite:
 ```plaintext
+@misc{wang2025knowbettermodelinghumanlike,
+      title={Know You First and Be You Better: Modeling Human-Like User Simulators via Implicit Profiles},
+      author={Kuang Wang and Xianfei Li and Shenghao Yang and Li Zhou and Feng Jiang and Haizhou Li},
+      year={2025},
+      eprint={2502.18968},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL},
+      url={https://arxiv.org/abs/2502.18968},
+}
 ```