---
license: apache-2.0
language:
- ko
- en
pipeline_tag: text-generation
library_name: transformers
base_model: PoSTMEDIA/Rosetta-7B-Base
tags:
- rosetta
- korean
- bilingual
- postmedia
---
# Rosetta-7B-Instruct
[](https://huggingface.co/collections/PoSTMEDIA/rosetta-6a9db30fd1b4585b0c1845e9)
[](https://github.com/PoSTMEDIA-AI/vllm)
[](./LICENSE)
## Introduction
**Rosetta-7B-Instruct** is a 7B-parameter bilingual (Korean-English) instruction-tuned model developed by **PoSTMEDIA**. Built on PoSTMEDIA's **Rosetta** dense decoder-only architecture, it is post-trained from [Rosetta-7B-Base](https://huggingface.co/PoSTMEDIA/Rosetta-7B-Base) through large-scale supervised fine-tuning and preference optimization, with a deliberate focus on the capabilities that matter most in real Korean-language deployments: instruction following, Korean cultural and factual knowledge, and Korean mathematical reasoning.
Rosetta-7B is a three-model family covering the full spectrum from research to production:
| Model | Download | Note |
|---|---|---|
| Rosetta-7B-Base | [HuggingFace](https://huggingface.co/PoSTMEDIA/Rosetta-7B-Base) | Foundation model (completion-style) |
| **Rosetta-7B-Instruct** | [HuggingFace](https://huggingface.co/PoSTMEDIA/Rosetta-7B-Instruct) | Instruction following / chat *(this model)* |
| Rosetta-7B-Instruct-NVFP4 | [HuggingFace](https://huggingface.co/PoSTMEDIA/Rosetta-7B-Instruct-NVFP4) | NVFP4 4-bit of Instruct (NVIDIA Blackwell / DGX Spark) |
| Rosetta-7B-Think | [HuggingFace](https://huggingface.co/PoSTMEDIA/Rosetta-7B-Think) | Explicit reasoning (``) |
| Rosetta-7B-Think-NVFP4 | [HuggingFace](https://huggingface.co/PoSTMEDIA/Rosetta-7B-Think-NVFP4) | NVFP4 4-bit of Think (NVIDIA Blackwell / DGX Spark) |
## Highlights
- **Korean math reasoning leader** — HRM8K **60.4**, the top score in our eight-model same-protocol comparison of Korean and global open models
- **Top-tier instruction following** — IFEval **80.4**, within the leading group of Korean open models
- **Korean-first design** — 161K Korean-extended vocabulary, dedicated Korean continual-pretraining stage, and in-house synthetic Korean data assets
- **Strong Korean factuality** — KoSimpleQA **71.7**
- **65,536-token context window** with interleaved local-global attention
- **Apache-2.0** — unrestricted commercial use
## Model Summary
| Architecture | Rosetta dense decoder-only Transformer (RosettaForCausalLM) |
| Parameters | 7B |
| Layers | 32 |
| Hidden size | 4,096 |
| Attention heads | 32 |
| Attention pattern | interleaved sliding-window (4,096) + global, 3:1, with QK-normalization |
| Context length | 65,536 |
| Vocabulary | 161,425 (Korean-extended) |
| Post-training | SFT → preference optimization (DPO) |
| License | Apache-2.0 |
## Training Overview
The Rosetta-7B family was built through a multi-stage pipeline carried out end-to-end by PoSTMEDIA:
1. **Pretraining** on trillions of tokens of curated bilingual web, code, and academic text
2. **Staged mid-training** for reasoning-dense data and long-context extension up to 65K
3. **Korean continual pretraining** on curated Korean corpora plus in-house synthetic Korean data assets, with the vocabulary extended to 161K for efficient Korean tokenization
4. **Post-training** — large-scale supervised fine-tuning followed by preference optimization (DPO) targeting instruction following, factuality, and safe refusals
## Evaluation Results
All models in the table below, **including competitors, were re-evaluated in-house under an identical protocol** (lm-evaluation-harness + vLLM ≥ 0.26, identical prompts, decoding parameters, and generation budgets). Scores are therefore directly comparable within this table, but may differ from numbers reported elsewhere under different setups.
| Benchmark |
Rosetta-7B-Instruct
7B |
A.X-4.0-Light
7B |
A.X-3.1-Light
7B |
EXAONE-3.5
7.8B |
Midm-2.0-Base
11.5B |
kanana-2
3B |
Llama-3.1
8B |
Ministral-3
8B |
| General & Reasoning |
| MMLU |
60.2 |
73.0 |
57.7 |
73.4 |
68.0 |
59.5 |
66.6 |
72.4 |
| GPQA |
39.4 |
37.4 |
30.8 |
36.9 |
31.8 |
32.3 |
22.7 |
49.5 |
| GSM8K |
79.8 |
68.1 |
84.0 |
88.9 |
80.7 |
70.0 |
84.2 |
80.8 |
| IFEval |
80.4 |
82.8 |
77.3 |
82.1 |
81.3 |
76.5 |
80.0 |
63.8 |
| Korean Language & Knowledge |
| KMMLU |
45.6 |
58.1 |
49.9 |
52.0 |
56.4 |
44.4 |
42.7 |
48.5 |
| CLIcK |
59.2 |
68.1 |
53.0 |
66.2 |
70.7 |
58.2 |
54.7 |
55.0 |
| HAE-RAE |
63.0 |
74.0 |
67.0 |
73.7 |
79.2 |
70.8 |
55.3 |
56.8 |
| HRM8K |
60.4 |
46.4 |
51.1 |
49.8 |
51.2 |
31.4 |
30.8 |
54.2 |
| KoSimpleQA† |
71.7 |
77.5 |
45.9 |
78.9 |
88.9 |
81.0 |
— |
72.1 |
Bold indicates the best score in each row. — indicates not evaluated. † KoSimpleQA is evaluated as a judge-free 10-choice MCQA variant. Generative benchmarks use greedy decoding; subtask-style suites report the unweighted subtask mean.
In the broadest same-protocol comparison we are aware of for Korean open models, Rosetta-7B-Instruct holds the **top score on Korean mathematical reasoning (HRM8K 60.4)** — ahead of every Korean flagship in the table — while placing in the leading group on instruction following (IFEval 80.4) and Korean factuality (KoSimpleQA 71.7). Models that lead on knowledge-recall benchmarks are either substantially larger (Midm-2.0-Base, 11.5B) or newest-generation flagships, and none of them match Rosetta on Korean math.
## Quickstart
### Transformers
Requires `transformers>=5.13` and `trust_remote_code=True` (the Rosetta architecture ships as custom code in this repository).
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "PoSTMEDIA/Rosetta-7B-Instruct"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id, dtype="bfloat16", device_map="auto", trust_remote_code=True
)
messages = [
{"role": "user", "content": "한국의 전통 발효 음식 세 가지를 소개해줘."},
]
inputs = tokenizer.apply_chat_template(
messages, add_generation_prompt=True, return_tensors="pt"
).to(model.device)
out = model.generate(inputs, max_new_tokens=1024, temperature=0.7, top_p=0.9, do_sample=True)
print(tokenizer.decode(out[0][inputs.shape[1]:], skip_special_tokens=True))
```
### vLLM
Use the PoSTMEDIA vLLM distribution with native Rosetta support — no `trust_remote_code` required:
```bash
VLLM_USE_PRECOMPILED=1 pip install git+https://github.com/PoSTMEDIA-AI/vllm@rosetta-v0.26.0
vllm serve PoSTMEDIA/Rosetta-7B-Instruct --dtype bfloat16
```
```python
from openai import OpenAI
client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")
resp = client.chat.completions.create(
model="PoSTMEDIA/Rosetta-7B-Instruct",
messages=[{"role": "user", "content": "부산 여행 1박 2일 코스를 짜줘."}],
temperature=0.7,
)
print(resp.choices[0].message.content)
```
> [!IMPORTANT]
> vLLM **v0.26 or later** is required. Recommended sampling: `temperature 0.7, top_p 0.9` (or greedy for deterministic tasks).
## Limitations
- The model can generate factually incorrect or outdated information; verify outputs for high-stakes use.
- Outputs may reflect biases present in web-scale training data.
- Optimized for Korean and English; other languages are not guaranteed.
- Alignment was performed on contexts up to 32K tokens; validate quality for longer inputs.
## License
Apache License 2.0 — see [LICENSE](./LICENSE). If you build something with Rosetta, we'd appreciate a *"Built with Rosetta"* attribution.
## Citation
```bibtex
@misc{rosetta2026,
title = {Rosetta-7B: A Bilingual Korean-English Language Model Family},
author = {{PoSTMEDIA AI Lab}},
year = {2026},
url = {https://huggingface.co/collections/PoSTMEDIA/rosetta-6a9db30fd1b4585b0c1845e9}
}
```
## Contact
Questions and feedback — please open a discussion on the model page.