W-Train Utils for MLflow Triton Plugin
Project description
w-train-utils-mlflow-triton-plugin
가상환경 설정
pyenv install 3.8.18
pyenv virtualenv 3.8.18 wtrainclient3.8
pyenv activate wtrainclient3.8
Triton Inference Server 실행
$ docker run --rm -p8000:8000 -p8001:8001 -p8002:8002 \
-e AWS_ACCESS_KEY_ID=<AccessKey> \
-e AWS_SECRET_ACCESS_KEY=<SecretKey> \
nvcr.io/nvidia/tritonserver:24.01-py3 \
tritonserver --model-repository=s3://https://kitech-minio-api.wimcorp.dev:443/triton \
--model-control-mode=explicit \
--log-verbose=1
환경 변수 설정
프로젝트를 실행하기 전에 아래의 환경 변수들을 설정해야 합니다:
환경변수 | 설명 | 예시 |
---|---|---|
MLFLOW_S3_ENDPOINT_URL | MLflow가 저장소로 사용하고있는 MinIO 엔드포인트 URL | http://localhost:9000 |
MLFLOW_TRACKING_URI | MLflow 트래킹 서버의 URI | http://localhost:5001 |
AWS_ACCESS_KEY_ID | MinIO 서버 접근을 위한 AWS 호환 액세스 키 | minio |
AWS_SECRET_ACCESS_KEY | MinIO 서버 접근을 위한 AWS 호환 시크릿 액세스 키 | miniostorage |
TRITON_URL | Triton Inference Server 의 grpc 엔드포인트 URL | http://localhost:8001 |
TRITON_MODEL_REPO | Triton Inference Server 의 모델저장소 URL | s3://http://localhost:9000/triton |
패키지 빌드 및 업로드
# 필요한 의존성 설치
pip install wheel setuptools twine
vi ~/.pypirc
[distutils]
index-servers =
pypi
pypi-repository
[pypi]
username = __token__
password = <token>
[pypi-repository]
repository: https://<domain>/repository/<pypi-hosted>/
username: <username>
password: <password>
sh scripts/build.sh
sh scripts/deploy.sh
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file wtu-mlflow-triton-plugin-0.0.15.tar.gz
.
File metadata
- Download URL: wtu-mlflow-triton-plugin-0.0.15.tar.gz
- Upload date:
- Size: 11.7 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.10.12
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | e6d1db9bc3de31f8ee2ef0ec2e612c92f4be6a9cb41da39de5d6b6840693db59 |
|
MD5 | 95e28079f3abd8a3b75535e74127132b |
|
BLAKE2b-256 | 624bcbed4f01ae5d92ed2ec8d48cafae8ca8fa76a64fe08f4a7a1ee60fc92821 |
File details
Details for the file wtu_mlflow_triton_plugin-0.0.15-py3-none-any.whl
.
File metadata
- Download URL: wtu_mlflow_triton_plugin-0.0.15-py3-none-any.whl
- Upload date:
- Size: 13.3 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/5.1.1 CPython/3.10.12
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 912b3339a6613500bb1f00ca7ac3bedd46e1bab20ada915a316fa0cd8399981f |
|
MD5 | 1cbd2873e6236e4842566c7a511580f7 |
|
BLAKE2b-256 | a4480f2be6b932e72f2a02fb74a6914b3fd30cc53cb5e680a6bb6686144e718b |