Instructions to use liushiliushi/ConfTuner-Ministral with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use liushiliushi/ConfTuner-Ministral with Transformers:

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

pipe = pipeline("text-generation", model="liushiliushi/ConfTuner-Ministral")
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("liushiliushi/ConfTuner-Ministral")
model = AutoModelForCausalLM.from_pretrained("liushiliushi/ConfTuner-Ministral", 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]:]))

PEFT
How to use liushiliushi/ConfTuner-Ministral with PEFT:
```
Task type is invalid.
```
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use liushiliushi/ConfTuner-Ministral with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/liushiliushi/ConfTuner-Ministral

SGLang

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

Docker Model Runner
How to use liushiliushi/ConfTuner-Ministral with Docker Model Runner:
```
docker model run hf.co/liushiliushi/ConfTuner-Ministral
```

Improve model card: Add pipeline tag, library, description, and usage instructions

by nielsr HF Staff - opened Sep 19, 2025

base: refs/heads/main

←

from: refs/pr/1

Discussion Files changed

+101

-132

nielsr

Sep 19, 2025

This PR significantly improves the model card for the ConfTuner-Ministral model by:

Updating the model card's main title to reflect the paper: "ConfTuner: Training Large Language Models to Express Their Confidence Verbally".
Providing a comprehensive description based on the paper's abstract, filling in Model Details such as developed by, model type, and language.
Adding the pipeline_tag: text-generation to ensure the model is correctly categorized and discoverable on the Hugging Face Hub.
Setting library_name: transformers to indicate compatibility with the transformers library for inference, which is a standard for PEFT adapters based on Mistral.
Specifying license: other in the metadata and adding a note in the content, as no explicit license for the fine-tuned model weights was found.
Adding direct links to the paper (https://huggingface.co/papers/2508.18847) and the GitHub repository (https://github.com/liushiliushi/ConfTuner) for easy access to further information and code.
Populating the How to Get Started with the Model section with instructions on using the inference.py script, adhering strictly to the disclaimer of not creating code snippets not present in the original GitHub README.
Including detailed Training Details and Evaluation sections based on the GitHub README and paper.
Adding a correct BibTeX citation for the ConfTuner paper.
Updating Framework versions with PEFT and Transformers information.

Improve model card: Add pipeline tag, library, description, and usage instructions23fb0d46

liushiliushi changed pull request status to merged Sep 20, 2025

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment