# 4.3.1 Hosting YOLOv8 PyTorch Model

1. Amazon S3에 yolov8-'name'으로 버킷을 생성합니다.

<figure><img src="/files/li8iZ74CzRDfJMED0b9u" alt=""><figcaption></figcaption></figure>

2. 이전 Lab에서 활용했던 SageMaker Studio에 접속합니다.
3. 노트북 환경을 설정합니다. 배포하는 컨테이너 이미지가 PyTorch 1.12버전 이므로 PyTorch 1.12 Python 3.8 GPU Optimized image를 선택합니다.

<figure><img src="/files/SW3bkKSrg20Me80acB07" alt=""><figcaption></figcaption></figure>

4. 아래 소스코드를 다운로드 합니다.

{% file src="/files/LmqLAvOzFR8ll5bpS3fN" %}

5. 업로드 버튼을 클릭하여 압축파일을 업로드합니다.

<figure><img src="/files/FOvcgj19xJFML7K5WKCe" alt=""><figcaption></figcaption></figure>

6. File - New - Terminal 을 클릭하고, 아래 코드를 입력합니다.

```
sudo yum install unzip
unzip sm-notebook.zip
```

7. 압축이 해제되면 1\_DeployEndpoint.ipynb 파일을 더블클릭하여 노트북을 엽니다.
8. 1\_DeployEndpoint.ipynb 노트북은 먼저 Ultralytics에서 제공하는 Pretrained 모델을 다운로드하게 됩니다. Pose model을 테스트해야 하므로, 2번째 블록에 있는 model\_name을 'yolov8n-pose.pt'로 변경합니다.

<pre><code>!pip3 install ultralytics
from ultralytics import YOLO

## Choose a model:
model_name = <a data-footnote-ref href="#user-content-fn-1">'yolov8n-pose.pt'</a> #replace pose model

YOLO(model_name)
os.system(f'mv {model_name} code/.')

bashCommand = "tar -cpzf  model.tar.gz code/"
process = subprocess.Popen(bashCommand.split(), stdout=subprocess.PIPE)
output, error = process.communicate()
</code></pre>

9. 노트북 내 Deploy the model on SageMaker Endpoint 블록에 'INSTANCE\_TYPE'을 선택합니다.

```
INSTANCE_TYPE = 'ml.m5.4xlarge'
ENDPOINT_NAME = 'yolov8-pytorch-' + str(datetime.utcnow().strftime('%Y-%m-%d-%H-%M-%S-%f'))

# Store the endpoint name in the history to be accessed by 2_TestEndpoint.ipynb notebook
%store ENDPOINT_NAME
print(f'Endpoint Name: {ENDPOINT_NAME}')

predictor = model.deploy(initial_instance_count=1, 
                         instance_type=INSTANCE_TYPE,
                         deserializer=JSONDeserializer(),
                         endpoint_name=ENDPOINT_NAME)
```

10. SageMaker Studio 메뉴 탭의 Run - Run All Cells를 클릭하면 노트북 내 모든 셀이 실행 되며 YOLOv8 모델이 SageMaker Endpoint에 배포됩니다.

<figure><img src="/files/pZU8i6aQGoYhbTTv9MmO" alt=""><figcaption></figcaption></figure>

11. Home 탭 - Deployments - Endpoints 에 들어가면 배포된 YOLOv8 모델의 Endpoint를 확인할 수 있습니다.

<figure><img src="/files/0FkY7s6Ap42FxsMezGPE" alt=""><figcaption></figcaption></figure>

[^1]:


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://cheolmin-ki.gitbook.io/sagemaker-image-workshop/lab4.-use-your-own-model-or-algorithm/4.3-deploy-sagemaker-endpoint-with-pretrained-model/4.3.1-hosting-yolov8-pytorch-model.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.