Instructions to use ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2") 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/Qwen3.8-27B-Human-KO-Enterprise-v0.2") model = AutoModelForMultimodalLM.from_pretrained("ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2", 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/Qwen3.8-27B-Human-KO-Enterprise-v0.2 with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2" # 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/Qwen3.8-27B-Human-KO-Enterprise-v0.2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2
- SGLang
How to use ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2 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/Qwen3.8-27B-Human-KO-Enterprise-v0.2" \ --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/Qwen3.8-27B-Human-KO-Enterprise-v0.2", "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/Qwen3.8-27B-Human-KO-Enterprise-v0.2" \ --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/Qwen3.8-27B-Human-KO-Enterprise-v0.2", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2 with Docker Model Runner:
docker model run hf.co/ThakiCloud/Qwen3.8-27B-Human-KO-Enterprise-v0.2
Qwen3.8-27B-Human-KO-Enterprise-v0.2
๋คํคํด๋ผ์ฐ๋(ThakiCloud)์ ํ๊ตญ์ด ๋ชจ๋ธ ThakiCloud/Qwen3.8-27B-Human-KO ๋ฅผ ๊ธฐ์
์๋ด ์์ด์ ํธ๋ก ์กฐ์ ํ ์ฒดํฌํฌ์ธํธ์
๋๋ค.
ํ์ฌ ์ ์ฑ
(๊ท์ ๊ณผ ์์ธ ์กฐํญ)๊ณผ ๊ณ ๊ฐ ์์ฒญ์ ํจ๊ป ์ฝ๊ณ , ๋ต๋ณยท๋๋ฌป๊ธฐยทํ์ธ ์์ฒญยท๋๊ตฌ ํธ์ถยท์๋ด์ฌ ์ฐ๊ฒฐยท๊ฑฐ์ ๊ฐ์ด๋ฐ
์ ์ฑ
์ ๋ง๋ ํ๋ ํ๋๋ฅผ JSON ์ผ๋ก ๊ณ ๋ฆ
๋๋ค.
๊ฒฐ๊ณผ โ ThakiCloud/EnterpriseOps-KO-Policy
| ๋ชจ๋ธ | ์ ํ๋ | ๊ฑฐ์ (REFUSE) | ๋๊ตฌ ํธ์ถ(CALL_TOOL) | ์์ธ ๋ฏธ๋ฐ๋ | ์์ธ ์์ |
|---|---|---|---|---|---|
ThakiCloud/Qwen3.8-27B-Human-KO (๊ธฐ์ค, 3ํ ํ๊ท ) |
95.6 | 97.1 | 79.4 | 90.3 | 97.2 |
Qwen3.8-27B-Human-KO-Enterprise-v0.1 |
97.2 | 97.1 | 80.5 | 96.6 | 99.4 |
Qwen3.8-27B-Human-KO-Enterprise-v0.2 |
97.6 | 99.2 | 79.2 | 94.9 | 96.4 |
๊ฐ์ ์กฐ๊ฑด์ผ๋ก ํ์ตํ ์ธ ์๋ ์ค ๊ฐ์ด๋ฐ ์ฑ์ ์ ์ฒดํฌํฌ์ธํธ๋ฅผ ๊ณต๊ฐํฉ๋๋ค(์ธ ์๋ ์ ํ๋ 97.1~97.6%, ํ๊ท 97.4%). ๊ธฐ์ค ๋ชจ๋ธ ๋๋น ์ค๋ต์ด 44.7% ์ค์์ต๋๋ค.
์์ ยท์ผ๋ฐ ๋ฅ๋ ฅ (๊ธฐ์ค ๋ชจ๋ธ๊ณผ ๊ฐ์ ์กฐ๊ฑด์์ ๋๋ํ ์ธก์ )
| ํญ๋ชฉ | ๊ธฐ์ค ๋ชจ๋ธ | ์ด ๋ชจ๋ธ | ๋ฌธํญ ์ |
|---|---|---|---|
| ์์ด ์ง์ (MMLU) | 92.5 | 95.7 | 94 |
| ํ๊ตญ์ด ์ง์ (KMMLU) | 63.0 | 68.0 | 100 |
| ์ฝ๋ฉ (HumanEval) | 91.1 | 95.6 | 90 |
| ์ง์ ๋ฐ๋ฅด๊ธฐ | 82.0 | 82.0 | 100 |
| ์ ํด ์์ฒญ ๊ฑฐ์ | 25/25 | 25/25 | 25 |
| ์ ํดํด ๋ณด์ด์ง๋ง ์ ์์ธ ์์ฒญ์ ๊ฑฐ์ ํ ์ | 0/25 | 2/25 | 25 |
๋ฌธํญ ์๊ฐ ์ ์ด 1~5%p ์ฐจ์ด๋ ํต๊ณ์ ์ผ๋ก ๊ตฌ๋ถ๋์ง ์์ต๋๋ค. ์ ํด ์์ฒญ์ ๋ ๋ชจ๋ธ ๋ชจ๋ ์ ๋ถ ๊ฑฐ์ ํ์ต๋๋ค.
์ฌ์ฉ๋ฒ
๋ฒค์น ์ ์ฅ์์ prompt_contract.txt ์ง์๋ฌธ์ ์์คํ
๋๋ ์ฌ์ฉ์ ๋ฉ์์ง ์์ ๋๊ณ , ์ ์ฑ
๊ณผ ๊ณ ๊ฐ ์์ฒญ์ ์ด์ด ๋ถ์
๋๋ค.
์๋ต์ {"action": ...} JSON ํ ๊ฐ์
๋๋ค. vLLM ๋ฑ Qwen3.8 ํธํ ์๋น์์ ๊ทธ๋๋ก ๋์ํฉ๋๋ค.
ํ์ต
LoRA ์ง๋ ๋ฏธ์ธ์กฐ์ ํ ๋ณํฉ. ํ์ต ๋ฐ์ดํฐ๋ ํฉ์ฑ ํ๊ตญ์ด ๊ธฐ์ ์ ์ฑ ์๋๋ฆฌ์ค์ด๋ฉฐ ๊ณต๊ฐํ์ง ์์ต๋๋ค. ๋ฒค์น ๋ฌธํญ์ ํ์ต์ ์ฐ์ง ์์์ต๋๋ค.
ํ๊ณ
๋ฒค์น๋ ํฉ์ฑ ์ ์ฑ ยท์์ฒญ์ผ๋ก ๋ง๋ค์์ต๋๋ค. ์ค์ ์ ๋ฌด์ ์ฐ๊ธฐ ์ ์ ์์ฌ ์ ์ฑ ๊ณผ ๋๊ตฌ๋ก ๋ฐ๋ก ๊ฒ์ฆํ์ธ์. ๋๊ตฌ ๋ชฉ๋ก์ ์ด ์ฒดํฌํฌ์ธํธ์ ๋ด์ฅ๋ผ ์์ง ์์ต๋๋ค โ ๋๊ตฌ ํธ์ถ์ด ํ์ํ ํ๋์ ๊ณ ๋ฅผ ๋ฟ, ์ค์ ํธ์ถ ์ธ์๋ ์๋น์ค ์ชฝ ์คํค๋ง๋ก ๋ฐ์์ผ ํฉ๋๋ค.
- Downloads last month
- 379