Training utilities for keynet - MLflow and model training support
Project description
keynet-train
MLflow와 통합된 모델 훈련 유틸리티
설치
pip install keynet-train
주요 기능
🚀 완전 자동화된 훈련 API
- 모델에서 자동으로 스키마 추론
- PyTorch 모델을 ONNX로 자동 변환
- MLflow에 자동 로깅 및 버전 관리
📊 지원 프레임워크
- PyTorch (TorchScript, ONNX 변환)
- ONNX (네이티브 지원)
- 다중 입력/출력 모델 지원
🔧 MLflow 통합
- 실험 자동 생성 및 관리
- 모델 아티팩트 자동 저장
- 메트릭 및 파라미터 추적
사용 예제
자동화된 PyTorch 훈련
from keynet_train import pytorch_trace_auto
import torch
# 단일 입력 모델
@pytorch_trace_auto("my_experiment", torch.randn(1, 3, 224, 224))
def train_simple_model():
model = MyModel()
# 훈련 로직
optimizer = torch.optim.Adam(model.parameters())
for epoch in range(10):
# ... 훈련 코드 ...
pass
return model # 모델만 반환!
# 다중 입력 모델
@pytorch_trace_auto("multi_input_exp", {
"image": torch.randn(1, 3, 224, 224),
"mask": torch.randn(1, 1, 224, 224)
})
def train_multi_input():
model = MultiInputModel()
# ... 훈련 ...
return model
라이선스
MIT License
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
keynet_train-0.0.1.dev0.tar.gz
(14.2 kB
view details)
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file keynet_train-0.0.1.dev0.tar.gz.
File metadata
- Download URL: keynet_train-0.0.1.dev0.tar.gz
- Upload date:
- Size: 14.2 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.12.10
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
a8864179d8ac014b34e958946c83d00ae4a4dec850d2b03f48dddbb8a8c89f55
|
|
| MD5 |
02cd8726ace495a2086b64c8f0db92c6
|
|
| BLAKE2b-256 |
f42f5992ff5eedf5712ec7455cdc29b3302cc4861325b6c9bacf488c3c752a20
|
File details
Details for the file keynet_train-0.0.1.dev0-py3-none-any.whl.
File metadata
- Download URL: keynet_train-0.0.1.dev0-py3-none-any.whl
- Upload date:
- Size: 18.0 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/6.1.0 CPython/3.12.10
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
33858c2463306da2d59f10c197d17ac138d586e0a736673f7146f825b54ce44c
|
|
| MD5 |
795b51bd918f941f1d720b6015784c25
|
|
| BLAKE2b-256 |
94e5f239a899841b49d4d0620281bea4dbca051a6e252414151376d8b3f4e621
|