File size: 14,884 Bytes
855e324 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 |
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
ko-gemma-2-9b-it - bnb 4bits
- Model creator: https://huggingface.co/rtzr/
- Original model: https://huggingface.co/rtzr/ko-gemma-2-9b-it/
Original model description:
---
license: gemma
library_name: transformers
pipeline_tag: text-generation
extra_gated_heading: Access Gemma on Hugging Face
extra_gated_prompt: >-
To access Gemma on Hugging Face, youโre required to review and agree to
Googleโs usage license. To do this, please ensure youโre logged in to Hugging
Face and click below. Requests are processed immediately.
extra_gated_button_content: Acknowledge license
tags:
- conversational
base_model:
- google/gemma-2-9b
language:
- ko
---
## Model Details
### Ko-Gemma-2-9B-IT
**[Ko-Gemma-2-9B-IT](https://huggingface.co/rtzr/ko-gemma-2-9b-it)** is a Korean-language conversational model that is part of the Gemma family of models. It is a text-to-text, decoder-only large language model, available in Korean. We fine-tuned this model on a carefully curated high-quality dataset using Supervised Fine-Tuning (SFT). And we use [Direct Preference Optimization](https://arxiv.org/abs/2305.18290) training specifically for Human Feedback. The datasets include:
- [Orca-Math](https://huggingface.co/datasets/kuotient/orca-math-korean-dpo-pairs)
- [dpo-mix-7k](https://huggingface.co/datasets/argilla/dpo-mix-7k)
Some of these datasets were partially used and translated for training. In particular, a lot of repetition occurred during the translation process, so preprocessing was performed based on N-gram.
#### *Inputs and outputs*
- **Input:** Text string, such as a question, a prompt, or a document to be summarized.
- **Output:** Generated Korean-language text in response to the input, such as an answer to a question, or a summary of a document.
### Google Gemma 2
Gemma is a family of lightweight, state-of-the-art open models from Google,
built from the same research and technology used to create the Gemini models.
They are text-to-text, decoder-only large language models, available in English,
with open weights for both pre-trained variants and instruction-tuned variants.
Gemma models are well-suited for a variety of text generation tasks, including
question answering, summarization, and reasoning. Their relatively small size
makes it possible to deploy them in environments with limited resources such as
a laptop, desktop or your own cloud infrastructure, democratizing access to
state of the art AI models and helping foster innovation for everyone.
## Benchmark Scores
We evaluated it internally using [LogicKor](https://github.com/instructkr/LogicKor) code. While the public LogicKor code is assessed as GPT-4, our internal evaluation was conducted as GPT-4o. Public scores will be added as they are released. The scores below include only 0-shot evaluations.
| Model | Math | Reasoning | Writing | Coding | Understanding | Grammar | Single ALL | Multi ALL | Overall |
|:---------:|:-----:|:------:|:-----:|:-----:|:----:|:-----:|:-----:|:-----:|:----:|
| [rtzr/ko-gemma-2-9b-it](https://huggingface.co/rtzr/ko-gemma-2-9b-it) | 8.71 / 8.00 | 9.14 / 8.00 | 9.43 / 9.29 | 9.00 / 9.43 | 9.57 / 9.86 | 7.14 / 5.00 | 8.83 | 8.26 | 8.55 |
| [google/gemma-2-9b-it](https://huggingface.co/google/gemma-2-9b-it) | 8.57 / 7.71 | 8.86 / 7.00 | 9.29 / 9.29 | 9.29 / 9.57 | 8.57 / 8.29 | 6.86 / 3.86 | 8.57 | 7.62 | 8.10 |
| [MLP-KTLim/llama-3-Korean-Bllossom-8B](https://huggingface.co/MLP-KTLim/llama-3-Korean-Bllossom-8B) | 6.43 / 5.71 | 6.86 / 5.14 | 9.14 / 8.57 | 8.29 / 8.14 | 8.43 / 9.29 | 5.71 / 5.29 | 7.48 | 7.02 | 7.25 |
| [yanolja/EEVE-Korean-Instruct-10.8B-v1.0](https://huggingface.co/yanolja/EEVE-Korean-Instruct-10.8B-v1.0) | 5.57 / 4.29 | 8.14 / 5.14 | 8.29 / 6.29 | 6.43 / 7.86 | 9.29 / 8.57 | 6.57 / 3.71 | 7.38 | 5.98 | 6.68 |
| [allganize/Llama-3-Alpha-Ko-8B-Instruct](https://huggingface.co/allganize/Llama-3-Alpha-Ko-8B-Instruct) | 4.57 / 3.00 | 6.86 / 6.43 | 7.43 / 6.71 | 8.43 / 8.43| 7.71 / 8.71 | 6.71 / 4.43 | 6.95 | 6.29 | 6.62 |
## Usage
### Install Dependencies
You must install transformers >= 4.42.3 for gemma2 models.
```bash
pip install transformers==4.42.3 accelerate
```
### Python code with Pipeline
```python
import transformers
import torch
model_id = "rtzr/ko-gemma-2-9b-it"
pipeline = transformers.pipeline(
"text-generation",
model=model_id,
model_kwargs={"torch_dtype": torch.bfloat16},
device_map="auto",
)
pipeline.model.eval()
instruction = "์์ธ์ ์ ๋ช
ํ ๊ด๊ด ์ฝ์ค๋ฅผ ๋ง๋ค์ด์ค๋?"
messages = [
{"role": "user", "content": f"{instruction}"}
]
prompt = pipeline.tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
terminators = [
pipeline.tokenizer.eos_token_id,
pipeline.tokenizer.convert_tokens_to_ids("<end_of_turn>")
]
outputs = pipeline(
prompt,
max_new_tokens=2048,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
print(outputs[0]["generated_text"][len(prompt):])
```
```markdown
์์ธ์ ์ญ์ฌ, ๋ฌธํ, ํ๋์ฑ์ด ์กฐํ๋ฅผ ์ด๋ฃฌ ๋งค๋ ฅ์ ์ธ ๋์์
๋๋ค. ์ฌ๊ธฐ์ ์ฆ๊ธธ ์ ์๋ ๋ค์ํ ๊ด๊ด์ง์ ๋ช
์๋ฅผ ์๊ฐํฉ๋๋ค. ๋ค์์ ์์ธ์ ์ ๋ช
ํ ๊ด๊ด ์ฝ์ค 3๊ฐ์ง์
๋๋ค.
**1. ์ญ์ฌ์ ๋ฌธํ๋ฅผ ๋๋ฌ์ผ ํ๊ตญ๊ด๊ด์ฝ์ค**
1. **๊ฒฝ๋ณต๊ถ**: ์กฐ์ ์๋์ ์
์ฅํ ์๊ถ์ ๋ง๋ฝํ ์ ์๋ ๊ณณ์
๋๋ค. ํนํ ๋งค๋
๋ด์ ์ด๋ฆฌ๋ '์ถ์ถ์ฐํ'๋ ๊ฒฝ๋ณต๊ถ์ ์๋ฆ๋ค์์ ๋์ฑ ๋๋ณด์ด๊ฒ ํฉ๋๋ค.
2. **๋ถ์ด ํ์ฅ๋ง์**: ๊ณ ํ์ค๋ฌ์ด ํ์ฅ์ด ๋ชจ์ฌ์๋ ๊ณณ์ผ๋ก, ์ ํต ๋ฌธํ ์ฒดํ์ด ๊ฐ๋ฅํฉ๋๋ค. '๋ถ์ด ํ์ฅ๋ง์ ๋ฌธํ์ฒดํ๊ด'์์๋ ํ๋ณต ์ฒดํ๋ถํฐ ์ข
์ด๋งํ, ํ๊ธ ์ฐ๊ธฐ ๋ฑ ๋ค์ํ ํ๋ก๊ทธ๋จ์ด ์ค๋น๋์ด ์์ต๋๋ค.
3. **์ธ์ฌ๋**: ์์ , ๋ฏธ์ ๊ด, ํ์๋น์ด ๋ง์ ๊ณณ์
๋๋ค. ํนํ '์ธ์ฌ๋ ๋ฌธํ๊ด'์์๋ ์์ธ์ ์ญ์ฌ์ ๋ฌธํ๋ฅผ ์ดํดํ๋ ๋ฐ ๋์์ด ๋๋ ์ ์๋ฅผ ๋ณผ ์ ์์ต๋๋ค.
4. **๊ดํ๋ฌธ** ๋ฐ **๋ช
๋**: ํ๋์ ์ธ ์ผํ๊ณผ ๋ ์คํ ๋์ด ์ฆ๋นํ ๊ณณ์
๋๋ค. ๊ดํ๋ฌธ์ ํนํ ์ ์์ด๋ค์ด ๋ง์ ๊ณณ์ผ๋ก, ์คํธ๋ฆฌํธ ํจ์
์ ๊ด์ฐฐํ๊ฑฐ๋ ๋ฐค๊ฑฐ๋ฆฌ์์ ํ๊ธฐ๋ฅผ ๋๋ ์ ์์ต๋๋ค.
**2. ๋์์ ๋ชจ์ต์ ๋ฐ๋ผ๋ณด๋ ๋ทฐํฌ์ด ์ฝ์ค**
1. **๋จ์ฐํ์**: ์์ธ์ ์์ง์ ์ธ ๊ฑด๋ฌผ๋ก, ๊ผญ๋๊ธฐ์์ ํผ์ณ์ง๋ 360๋์ ๊ฒฝ์น๊ฐ ์๋๋ค. ํนํ ๋ฐค์ด ๋๋ฉด ์กฐ๋ช
์ด ์ด์ฐ๋ฌ์ ธ ๋์ฑ ์๋ฆ๋ค์์ง๋๋ค.
2. **์์ธํ์**: ๋จ์ฐํ์์ ๋น์ทํ ์์น๋ก, ๋์ด๊ฐ ๋ ๋๊ธฐ ๋๋ฌธ์ ๋ ๋์ ์ ๋ง์ ๋ณผ ์ ์์ต๋๋ค. ์์ธํ์ ๋ด๋ถ์๋ ๋ค์ํ ์ ์๊ด๊ณผ ๋ ์คํ ๋๋ ์์ต๋๋ค.
3. **๋ถ์
์ฐ**: ์์ธ์ ์ค์ฌ๋ถ์ ์์นํ ์ฐ์ผ๋ก, ์์ธ์ ๊ฒฝ์น๋ฅผ ์กฐ๊ธ ๋ค๋ฅธ ๊ด์ ์์ ๋ณผ ์ ์์ต๋๋ค. ํนํ ๋ถ์
์ฐ ์ ์์ธ ๋ถ์
์ฌ์์๋ ์ข์ ์ ๋ง์ ๋ณผ ์ ์์ต๋๋ค.
4. **์์ธ์ฒ**: ๋
น์ง ๊ณต๊ฐ์ผ๋ก, ๋์์ ํผ์กํจ์์ ๋ฒ์ด๋ ์ ์๋ ๊ณณ์
๋๋ค. ๋ํ, ์์ธ์ฒ ๋ด๋ถ์๋ '์์ธ์ฒ ์ํธํ๋ ์ ํธ'๋ผ๋ ๊ณต๊ฐ์ด ์์ด ์์ ๊ณผ ์์ฐ์ ํจ๊ป ์ฒดํํ ์ ์์ต๋๋ค.
**3. ํ๋ ๋ฌธํ๋ฅผ ๋ง๋๋ ์ฝ์ค**
1. **์ผ์ฑ๋**: ํ๋ ๋ฏธ์ ๊ด์ด ๋ง์ ๊ณณ์ผ๋ก, '์ผ์ฑ ๋ฏธ์ ๊ด', '์๋ชจ๋ฆฌ์นด๋์ค ๊ฐค๋ฌ๋ฆฌ' ๋ฑ์ด ์์ต๋๋ค. ๋ํ, '์ฝ์์ค'๋ '์ํฌ์นด๋กํฌ์ค' ๋ฑ์ ๋ช
์๋ ๊ฐ๊น์ด ๊ณณ์ ์์ต๋๋ค.
2. **์ดํ์**: ์ธ๊ตญ์ธ๋ค์ด ๋ง์ ๊ณณ์ผ๋ก, ๋ค์ํ ์ธ๊ตญ ์์์ ์ฆ๊ธธ ์ ์๋ ๊ณณ์
๋๋ค. ๋ํ, '์ดํ์ ๊ธ๋ก์ปฌ๋ฌธํ์ผํฐ'์์๋ ์ธ๊ณ ๊ฐ๊ตญ์ ๋ฌธํ ์ฒดํ์ด ๊ฐ๋ฅํฉ๋๋ค.
3. **ํ๋**: ์ ์์ด๋ค์ ๋ฌธํ๊ฐ ๋์น๋ ๊ณณ์
๋๋ค. 'ํ๋ ๋กค๋งํ'์ ํนํ ๋ง์ ์ฌ๋๋ค์ด ๋ฐฉ๋ฌธํ๋ ๊ณณ์
๋๋ค. ๋ํ, 'ํ๋ ์์ ๊ฑฐ๋ฆฌ'์์๋ ๋
์์ ๋ฌธํ๋ฅผ ๋ง๋ ์ ์์ต๋๋ค.
4. **๊ฐ๋จ**: ์์ธ์ ํ๋์ ๋ชจ์ต์ ์ ๋ณด์ฌ์ฃผ๋ ๊ณณ์
๋๋ค. '๊ฐ๋จ์ญ'์ ์ค์ฌ์ผ๋ก ๋ง์ ๊ณ ๊ธ ์ผํ๋ชฐ๊ณผ ๋ ์คํ ๋์ด ์์ต๋๋ค.
์ด๋ฌํ ์ฝ์ค๋ฅผ ํตํด ์์ธ์ ๋ค์ํ ๋ชจ์ต์ ํ ๋ฒ์ ๋ง๋๋ณผ ์ ์์ ๊ฑฐ์์. ๊ฐ์์ ์ทจํฅ์ ๋ง์ถฐ ์ฝ์ค๋ฅผ ์กฐ์ ํ์๋ฉด ์ข๊ฒ ์ต๋๋ค. ์ฆ๊ฑฐ์ด ์ฌํ ๋์ธ์!
```
### Python code with AutoModel
```python
import os
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
model_id = "rtzr/ko-gemma-2-9b-it"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.bfloat16,
device_map="auto",
)
model.eval()
instruction = "์์ธ์ ์ ๋ช
ํ ๊ด๊ด ์ฝ์ค๋ฅผ ๋ง๋ค์ด์ค๋?"
messages = [
{"role": "user", "content": f"{instruction}"}
]
input_ids = tokenizer.apply_chat_template(
messages,
add_generation_prompt=True,
return_tensors="pt"
).to(model.device)
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids("<end_of_turn>")
]
outputs = model.generate(
input_ids,
max_new_tokens=2048,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
print(tokenizer.decode(outputs[0][input_ids.shape[-1]:], skip_special_tokens=True))
```
```markdown
์์ธ ๊ด๊ด ์ฝ์ค๋ฅผ ์ ์ํด๋๋ฆด๊ฒ์. ํ๋ฃจ ์ข
์ผ ์ฆ๊ฒ๊ฒ ์ฌํํ ์ ์๋ ๋ฃจํธ๋ก ๊ตฌ์ฑํ์ต๋๋ค.
### 1. ์์ธ์ญ์ฌ๊ด ๋ฐ ๋ถ์ดํ์ฅ๋ง์(์ค์ )
- ์์ธ์ญ์ฌ๊ด: ์์ธ์ ์ญ์ฌ์ ๋ฌธํ๋ฅผ ์ฒดํํ ์ ์๋ ๊ณณ์
๋๋ค. ๋ค์ํ ์ ์๋ฌผ๊ณผ ์์ค์ ์๋ฅผ ํตํด ์์ธ์ ๋ณํ๋ฅผ ์ดํด๋ณผ ์ ์์ต๋๋ค.
- ๋ถ์ดํ์ฅ๋ง์: ์์ธ์ ํ์ฅ์ ๋ณด์กดํ๊ณ ๊ด๋ฆฌํ๋ ๊ณณ์
๋๋ค. ์กฐ์ ์๋์ ๋ถ์๊ธฐ๋ฅผ ๋๋ ์ ์์ผ๋ฉฐ, ํ์ฅ์์ ๋ฌธํ ์ฝํ
์ธ ๋ ์ ๊ณตํ๋ ๊ณณ๋ ๋ง์ต๋๋ค.
### 2. ๋ถ์
์ฐ ์
์ฅ๊ณผ ๋ถ์
์ฐ ๋ฑ์ฐ(์ค์ )
- ๋ถ์
์ฐ์ ์์ธ์ ๋ถ์ชฝ์ ์์นํ ์ฐ์ผ๋ก, ์์ธ ํ๋ณตํ์์๋ ์์ฐ์ ๋ง๋ ์ ์๋ ๊ณณ์
๋๋ค. ๋ถ์
์ฐ ์
๊ตฌ์์ ๋ฑ์ฐ์ ์์ํ์ฌ, ๋ถ์
์ฐ ์ ์๊น์ง ์ฌ๋ผ๊ฐ๋ฉด ์์ธ์ ์ ๊ฒฝ์ ๋ณผ ์ ์์ต๋๋ค.
### 3. ์ข
๋ก ๋ช
๋ ์ผํ๊ณผ ๋ง์ง ํฌ์ด(๋ฎ)
- ๋ช
๋: ๋ค์ํ ์ผํ๋ชฐ๊ณผ ๋งค์ฅ์ด ์๋ ๊ณณ์
๋๋ค. ๋ช
๋ ์ผํํ์ด, ๋ฏธ์คํฐํธ์์คํฐ, ๋ฏธ์คํฐ๋ฆฌ๋ง์ผ ๋ฑ์ ๋ฐฉ๋ฌธํด๋ณด์ธ์.
- ๋ง์ง ํฌ์ด: ๋ช
๋์๋ ๋ค์ํ ์ง์ญ ์์์ ๋จน์ ์ ์๋ ๊ณณ์ด ๋ง์ต๋๋ค. ๋ก๋ณถ์ด, ์๋, ๋ญ๊ฐ์ ๋ฑ์ ๋ง๋ณผ ์ ์๋ ๊ณณ์ ์ถ์ฒ๋๋ฆฝ๋๋ค.
### 4. ์์ธ์๋ฆฝ๋ฏธ์ ๊ด๊ณผ ๋์๊ถ(์คํ)
- ์์ธ์๋ฆฝ๋ฏธ์ ๊ด: ํ๋๋ฏธ์ ์ ์ ์ํ๋ ๊ณณ์
๋๋ค. ํน๋ณ์ ์ด ์ด๋ฆฐ๋ค๋ฉด ๋ฐฉ๋ฌธํด ๋ณผ ์ ์์ต๋๋ค.
- ๋์๊ถ: ์กฐ์ ์๋์ ๊ถ๊ถ์
๋๋ค. ํนํ ๋ด์๋ ๋ฒ๊ฝ์ด ์๋ฆ๋ต๊ฒ ๋ง๋ฐํฉ๋๋ค.
### 5. ๋จ์ฐํ์์ ๋จ์ฐ๊ณต์ ์ฐ์ฑ
(์คํ)
- ๋จ์ฐํ์: ๋จ์ฐ์ ์๋ ๊ด๋๋์
๋๋ค. ๋จ์ฐํ์์ ์ฌ๋ผ๊ฐ๋ฉด ์์ธ์ 360๋ ์ ๊ฒฝ์ ๋ณผ ์ ์์ต๋๋ค.
- ๋จ์ฐ๊ณต์: ๋จ์ฐ์ ์๋ ๊ณต์์
๋๋ค. ๋ค์ํ ํ
๋ง ๊ณต์๊ณผ ์กฐ๊ฒฝ์ด ์ ๋ ๊ณณ์
๋๋ค. ๋จ์ฐ๊ณต์์ ์ฐ์ฑ
ํ๋ฉฐ ํด์์ ์ทจํ ์ ์์ต๋๋ค.
### 6. ๋ช
๋ ๋๋ ์ดํ์์์์ ์ ๋
์์ฌ์ ๋ฌธํ ํ๋(์ ๋
)
- ๋ช
๋: ๋ค์ํ ์ ํต์ ์ธ ํ๊ตญ ์์์ ๋จน์ ์ ์๋ ๊ณณ์
๋๋ค. ๋ํ, ๋ช
๋์ ๋ฐค์๋ ํ๊ธฐ์ฐจ๊ฒ ํ๋ฐํ ๋ฌธํ ์ํ์ ํ ์ ์๋ ๊ณณ์
๋๋ค.
- ์ดํ์: ์ธ๊ตญ์ธ ๊ด๊ด๊ฐ๋ค์ด ๋ง์ด ์ฐพ๋ ๊ณณ์ผ๋ก, ๋ค์ํ ์ธ๊ณ ์์์ ๋จน์ ์ ์์ผ๋ฉฐ, ํด๋ฝ์ด๋ ๋ฐ๊ฐ ๋ง์ ๋ฌธํ์ ํ๋์ด ๊ฐ๋ฅํ ๊ณณ์
๋๋ค.
์ด ์ฝ์ค๋ ํ๋ฃจ ์ข
์ผ ํ๋ฐํ๊ฒ ์ฌํ์ ํ ์ ์๋๋ก ๊ณํํ์ต๋๋ค. ๊ฐ ์ง์ญ์ ๋ฐ๋ผ ์ด๋ ์๊ฐ์ ๊ณ ๋ คํ์๊ณ , ๊ฐ์ฅ ์๊ฐ๊ณผ ์ ์ ์ผ์ ๋ฑ์ ๋ฏธ๋ฆฌ ํ์ธํ์๋ ๊ฒ์ด ์ข์ต๋๋ค. ์ฆ๊ฑฐ์ด ์ฌํ ๋์ธ์!
```
### Quantized Versions through bitsandbytes
- *Using 8-bit precision*
- *Using 4-bit precision*
```python
# pip install bitsandbytes
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
model_id = "rtzr/ko-gemma-2-9b-it"
quantization_config_8bit = BitsAndBytesConfig(load_in_8bit=True)
# quantization_config_4bit = BitsAndBytesConfig(load_in_4bit=True)
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.bfloat16,
device_map="auto",
quantization_config=quantization_config_8bit,
# quantization_config=quantization_config_4bit,
low_cpu_mem_usage=True,
)
model.eval()
instruction = "์์ธ์ ์ ๋ช
ํ ๊ด๊ด ์ฝ์ค๋ฅผ ๋ง๋ค์ด์ค๋?"
messages = [
{"role": "user", "content": f"{instruction}"}
]
input_ids = tokenizer.apply_chat_template(
messages,
add_generation_prompt=True,
return_tensors="pt"
).to(model.device)
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids("<end_of_turn>")
]
outputs = model.generate(
input_ids,
max_new_tokens=2048,
eos_token_id=terminators,
do_sample=True,
temperature=0.6,
top_p=0.9,
)
print(tokenizer.decode(outputs[0][input_ids.shape[-1]:], skip_special_tokens=True))
```
### VLLM Usage
When we use `vllm==0.5.1`, the gemma2 model cannot be loaded yet and the following [issue](https://github.com/vllm-project/vllm/issues/6237) occurs. So it is recommended to use `vllm/vllm-openai:latest` docker or [`vllm==0.5.0.post1`](https://github.com/vllm-project/vllm/releases/tag/v0.5.0.post1).
```bash
#!/bin/bash
VLLM_ATTENTION_BACKEND=FLASHINFER
MODEL_NAME="rtzr/ko-gemma-2-9b-it"
MODEL_PATH="YOUR_PATH/${MODEL_NAME}"
docker run --rm --gpus all \
-p 8000:8000 \
--shm-size=12gb --ulimit memlock=-1 --ulimit stack=67108864 \
-e VLLM_ATTENTION_BACKEND=${VLLM_ATTENTION_BACKEND} \
-v $MODEL_PATH:/vllm-workspace/${MODEL_NAME} \
vllm/vllm-openai:latest \
--model ${MODEL_NAME} --dtype auto \
--gpu-memory-utilization 0.8
```
## License
Gemma 2 License: <https://ai.google.dev/gemma/terms>
## Citation
```none
@article{RTZR,
title={ko-gemma-2-9b-it},
author={Return Zero Team},
year={2024},
url={https://huggingface.co/rtzr/ko-gemma-2-9b-it}
}
```
```none
@article{gemma_2024,
title={Gemma},
url={https://www.kaggle.com/m/3301},
DOI={10.34740/KAGGLE/M/3301},
publisher={Kaggle},
author={Gemma Team},
year={2024}
}
```
|