Instructions to use ThakiCloud/FactGuard-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use ThakiCloud/FactGuard-4B with Transformers:

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

pipe = pipeline("text-generation", model="ThakiCloud/FactGuard-4B")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("ThakiCloud/FactGuard-4B")
model = AutoModelForMultimodalLM.from_pretrained("ThakiCloud/FactGuard-4B", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use ThakiCloud/FactGuard-4B with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/ThakiCloud/FactGuard-4B

SGLang

How to use ThakiCloud/FactGuard-4B 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 "ThakiCloud/FactGuard-4B" \
    --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": "ThakiCloud/FactGuard-4B",
		"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 "ThakiCloud/FactGuard-4B" \
        --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": "ThakiCloud/FactGuard-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use ThakiCloud/FactGuard-4B with Docker Model Runner:
```
docker model run hf.co/ThakiCloud/FactGuard-4B
```

FactGuard-4B

File size: 1,685 Bytes

29a66c8

"""Minimal FactGuard-4B inference: returns label probabilities for one claim against evidence lines."""
import json, torch
from transformers import AutoTokenizer, AutoModelForCausalLM

REPO = "ThakiCloud/FactGuard-4B"
HEAD = ("공식 근거 사실(F1~F9)과 AI 답변의 주장 하나를 대조해, 보여 준 근거만으로 주장을 판정한다(조건부 함의).\n"
        "A=지지(근거가 주장 전체를 함의: 대상 제품·플랜·수치·조건·기간이 모두 근거에 있다) B=모순(근거가 주장과 양립할 수 없다고 명시) "
        "C=확인불가(근거가 말하지 않거나 플랜·열 귀속·조건을 확정할 수 없다. 빈 칸·흐린 칸·헤더와 안 맞는 행은 근거가 아니다) D=과거 근거로만 지지\n"
        "결정 근거: a=없음 b..j=F1..F9\n\n")
LABELS = ["SUPPORTED", "CONTRADICTED", "NOT_VERIFIABLE", "STALE"]

tok = AutoTokenizer.from_pretrained(REPO)
model = AutoModelForCausalLM.from_pretrained(REPO, dtype=torch.bfloat16, device_map="auto").eval()
label_ids = [tok.encode(" " + t, add_special_tokens=False)[0] for t in "ABCD"]

packet = ("대상 브랜드: ExampleSaaS\n근거:\n"
          "F1: [Pricing | ExampleSaaS] Pro — $20 / month, billed monthly\n"
          "F2: [Pricing | ExampleSaaS] Pro — Up to 10 projects\n"
          "주장: ExampleSaaS Pro 플랜은 월 결제 시 월 20달러이다.")
ids = tok.encode(HEAD + packet + "\n\n판정:", add_special_tokens=False)[-1024:]
with torch.no_grad():
    logits = model(torch.tensor([ids], device=model.device)).logits[0, -1, label_ids].float()
probs = torch.softmax(logits, -1).tolist()
print(dict(zip(LABELS, [round(p, 4) for p in probs])))