Text Generation
Transformers
Safetensors
qwen3_5
qwen3.5
korean
reasoning
thinking
sft
k-ai
conversational
Instructions to use Anserwise/AWAXIS-Think-27b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Anserwise/AWAXIS-Think-27b with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Anserwise/AWAXIS-Think-27b") messages = [ {"role": "user", "content": "Who are you?"}, ] pipe(messages)# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("Anserwise/AWAXIS-Think-27b") model = AutoModelForMultimodalLM.from_pretrained("Anserwise/AWAXIS-Think-27b") messages = [ {"role": "user", "content": "Who are you?"}, ] 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=40) print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Anserwise/AWAXIS-Think-27b with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Anserwise/AWAXIS-Think-27b" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Anserwise/AWAXIS-Think-27b", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }'Use Docker
docker model run hf.co/Anserwise/AWAXIS-Think-27b
- SGLang
How to use Anserwise/AWAXIS-Think-27b 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 "Anserwise/AWAXIS-Think-27b" \ --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": "Anserwise/AWAXIS-Think-27b", "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 "Anserwise/AWAXIS-Think-27b" \ --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": "Anserwise/AWAXIS-Think-27b", "messages": [ { "role": "user", "content": "What is the capital of France?" } ] }' - Docker Model Runner
How to use Anserwise/AWAXIS-Think-27b with Docker Model Runner:
docker model run hf.co/Anserwise/AWAXIS-Think-27b
File size: 3,673 Bytes
a1c5518 11d6c91 a1c5518 11d6c91 a1c5518 11d6c91 a1c5518 11d6c91 dff8872 11d6c91 dff8872 11d6c91 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 | ---
language:
- ko
- en
- ja
- zh
- multilingual
license: apache-2.0
tags:
- qwen3.5
- korean
- reasoning
- thinking
- sft
- k-ai
base_model:
- FINAL-Bench/Darwin-27B-Opus
pipeline_tag: text-generation
library_name: transformers
---
# AWAXIS-Think-27b
[FINAL-Bench/Darwin-27B-Opus](https://huggingface.co/FINAL-Bench/Darwin-27B-Opus) ๊ธฐ๋ฐ, ํ๊ตญ์ด ํนํ ๊ณ ํ์ง SFT๋ฅผ ์ํํ ์ถ๋ก ๋ชจ๋ธ์
๋๋ค.
> โ ๏ธ **Requirements / Loading ์ฃผ์**
> ์ด ๋ชจ๋ธ์ `model_type: qwen3_5_text` (Qwen3.5 ํ์ด๋ธ๋ฆฌ๋ ์ํคํ
์ฒ)๋ฅผ ์ฌ์ฉํฉ๋๋ค.
> **`transformers >= 5.5.4` ์ด์** ์์๋ง ์ ์ ๋ก๋๋ฉ๋๋ค.
>
> ```bash
> pip install --upgrade "transformers>=5.5.4"
> # ๋๋ ์ต์ ๊ฐ๋ฐํ
> pip install "transformers @ git+https://github.com/huggingface/transformers.git@main"
> ```
>
> ๊ตฌ๋ฒ์ transformers์์ ๋ํ๋๋ `model_type 'qwen3_5_text'๋ฅผ ์ธ์ํ์ง ๋ชปํจ` ์ค๋ฅ๋
> ๋ผ์ด๋ธ๋ฌ๋ฆฌ ๋ฏธ์
๋ฐ์ดํธ๋ก ์ธํ ๊ฒ์ด๋ฉฐ, ์ ๋ช
๋ น์ผ๋ก ํด๊ฒฐ๋ฉ๋๋ค.
## Method
- **Base Model**: [Darwin-27B-Opus](https://huggingface.co/FINAL-Bench/Darwin-27B-Opus) (Qwen3.5-27B family)
- **Korean SFT**: ํ๊ตญ์ด ๋ฌธํ, ์ญ์ฌ, ๋ฒ๋ฅ , ๊ฒฝ์ , ์ฌํ, ์ง๋ฆฌ ๋ฑ ํ๊ตญ ํนํ ์ง์ ์ค์ฌ์ ๊ณ ํ์ง instruction ๋ฐ์ดํฐ๋ก Supervised Fine-Tuning ์ํ
- **Thinking Mode**: `<think>` ํ๊ทธ๋ฅผ ํตํ Chain-of-Thought ๋จ๊ณ์ ์ถ๋ก ์ง์
## Benchmark
| Benchmark | Score |
|---|---|
| CLIcK (Korean Cultural & Linguistic Knowledge) | **81.0%** |
| KMMLU-Pro (Korean MMLU Professional) | **74.0%** |
## Model Specifications
| Property | Value |
|---|---|
| **Architecture** | Qwen3.5 Hybrid (GatedDeltaNet + Attention, 64 layers) |
| **Parameters** | ~27B |
| **Hidden Size** | 5120 |
| **Intermediate Size** | 16384 |
| **Context Length** | 262,144 tokens |
| **Precision** | BF16 |
| **Vocab Size** | 248,320 |
| **Thinking** | Supported (`<think>` tags) |
| **License** | Apache 2.0 |
## Usage
> **Requirements**: `transformers >= 5.5.4`
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
model = AutoModelForCausalLM.from_pretrained(
"Anserwise/AWAXIS-Think-27b",
torch_dtype=torch.bfloat16,
device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained("Anserwise/AWAXIS-Think-27b")
messages = [{"role": "user", "content": "์กฐ์ ์๋์ ๊ณผ๊ฑฐ์ ๋์ ๋ํด ์ค๋ช
ํด์ฃผ์ธ์."}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=1024, do_sample=False)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))
```
### vLLM
```bash
vllm serve Anserwise/AWAXIS-Think-27b \
--enforce-eager \
--max-model-len 32768 \
--dtype bfloat16
```
## Features
- Darwin-27B-Opus์ ๊ฐ๋ ฅํ ์ถ๋ก ๋ฅ๋ ฅ ๊ณ์น
- ํ๊ตญ์ด ๋ฌธํ, ์ญ์ฌ, ๋ฒ๋ฅ , ๊ฒฝ์ , ์ฌํ ๋ฑ ํ๊ตญ ํนํ ์ง์ ๊ฐํ
- Thinking mode๋ฅผ ํตํ ๋จ๊ณ์ ์ฌ๊ณ ์ถ๋ก
- ๋ค๊ตญ์ด ์ง์ (ํ๊ตญ์ด, ์์ด, ์ผ๋ณธ์ด, ์ค๊ตญ์ด)
- 262K ์ปจํ
์คํธ ๊ธธ์ด ์ง์
## Training
| Item | Details |
|---|---|
| **Base Model** | [FINAL-Bench/Darwin-27B-Opus](https://huggingface.co/FINAL-Bench/Darwin-27B-Opus) |
| **Method** | Korean-specialized Supervised Fine-Tuning |
| **Data** | ํ๊ตญ์ด ๋ฌธํยท์ง์ ์ค์ฌ ๊ณ ํ์ง instruction ๋ฐ์ดํฐ |
| **Developer** | [Anserwise](https://huggingface.co/Anserwise) |
## Acknowledgements
- [FINAL-Bench](https://huggingface.co/FINAL-Bench) โ Darwin-27B-Opus base model
- [Qwen Team](https://huggingface.co/Qwen) โ Qwen3.5 architecture
|