pytorch-lightning

KO | EN

gitlite — search

#python #java #python3 #arduino #golang #machine-learning #rust #html #flask #javascript #seismology #nodejs

pytorch-lightning

★ 31,288 Open GitHub ↗

코드 변경 없이 1개에서 10,000개 이상의 GPU에서 모든 크기의 AI 모델을 Pretrain 및 Finetune 하세요.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

**AI 모델의 사전 학습(Pretrain)과 미세 조정(Finetune)을 위한 딥러닝 프레임워크입니다.**

**모델을 서빙하시나요?** 순수 Python으로 커스텀 추론 서버를 구축하려면 [LitServe](https://github.com/Lightning-AI/litserve?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme)를 사용하세요.

______________________________________________________________________

<p align="center">
    <a href="#quick-start" style="margin: 0 10px;">Quick start</a> •
  <a href="#examples">Examples</a> •
  <a href="#why-pytorch-lightning">PyTorch Lightning</a> •
  <a href="#lightning-fabric-expert-control">Fabric</a> •
  <a href="https://lightning.ai/?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme">Lightning Cloud</a> •   
  <a href="#community">Community</a> •
  <a href="https://pytorch-lightning.readthedocs.io/en/stable/">Docs</a>
</p>

[![PyPI - Python Version](https://img.shields.io/pypi/pyversions/pytorch-lightning)](https://pypi.org/project/pytorch-lightning/)
[![PyPI Status](https://badge.fury.io/py/pytorch-lightning.svg)](https://badge.fury.io/py/pytorch-lightning)
[![PyPI - Downloads](https://img.shields.io/pypi/dm/pytorch-lightning)](https://pepy.tech/project/pytorch-lightning)
[![Conda](https://img.shields.io/conda/v/conda-forge/lightning?label=conda&color=success)](https://anaconda.org/conda-forge/lightning)
[![codecov](https://codecov.io/gh/Lightning-AI/pytorch-lightning/graph/badge.svg?token=SmzX8mnKlA)](https://codecov.io/gh/Lightning-AI/pytorch-lightning)

[![Discord](https://img.shields.io/discord/1077906959069626439?style=plastic)](https://discord.gg/VptPCZkGNa)
![GitHub commit activity](https://img.shields.io/github/commit-activity/w/lightning-ai/lightning)
[![license](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://github.com/Lightning-AI/pytorch-lightning/blob/master/LICENSE)

</div>

&nbsp;

</p>

</div>

&nbsp;

<a id="why-pytorch-lightning"></a>
# Why PyTorch Lightning?   
순수 PyTorch로 모델을 학습시키려면 많은 양의 반복적인 엔지니어링 코드를 작성하고 유지 관리해야 합니다. 역전파(backpropagation), 혼합 정밀도(mixed precision), multi-GPU 및 분산 학습을 처리하는 과정은 오류가 발생하기 쉬우며, 프로젝트마다 매번 다시 구현되는 경우가 많습니다. PyTorch Lightning은 PyTorch 코드를 체계화하여 이러한 인프라를 자동화하는 동시에, 모델 로직에 대한 완전한 제어권을 유지하게 해줍니다. 여러분은 과학(모델링)에 집중하세요. Lightning이 엔지니어링을 처리하며, 핵심 코드를 변경하지 않고도 CPU에서 multi-node GPU까지 확장합니다. PyTorch 전문가들은 여전히 [전문가 수준의 제어(expert-level control)](#lightning-fabric-expert-control) 기능을 선택할 수 있습니다.

재미있는 비유를 들자면: PyTorch가 Javascript라면, PyTorch Lightning은 ReactJS 또는 NextJS와 같습니다.

# Looking for GPUs?
[Lightning Cloud](https://lightning.ai/?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme)는 인프라 관리 없이 PyTorch Lightning을 실행하는 가장 쉬운 방법입니다. 명령어 하나로 학습을 시작하고 GPU, 오토스케일링, 모니터링 및 무료 티어 혜택을 누리세요. 클라우드 설정이 전혀 필요하지 않습니다.

또한, 본인의 하드웨어나 다른 클라우드 환경에서도 PyTorch Lightning을 실행할 수 있습니다.

# Lightning has 2 core packages

[PyTorch Lightning: PyTorch를 대규모로 학습 및 배포](#why-pytorch-lightning).
<br/>
[Lightning Fabric: 전문가 수준의 제어](#lightning-fabric-expert-control).

Lightning은 PyTorch 위에 어느 정도의 추상화를 추가할지에 대해 세밀한 제어 권한을 제공합니다.

&nbsp;

# Quick start
Lightning 설치:

```bash
pip install lightning
```

<details>
  <summary>Advanced install options</summary>

#### 선택적 의존성과 함께 설치

```bash
pip install lightning['extra']
```

#### Conda

```bash
conda install lightning -c conda-forge
```

#### 안정 버전 설치

소스에서 향후 릴리스 버전 설치

```bash
pip install https://github.com/Lightning-AI/lightning/archive/refs/heads/release/stable.zip -U
```

#### 최신 개발 버전(Bleeding-edge) 설치

소스에서 nightly 버전 설치 (보장되지 않음)

```bash
pip install https://github.com/Lightning-AI/lightning/archive/refs/heads/master.zip -U
```

또는 테스트용 PyPI에서 설치

```bash
pip install -iU https://test.pypi.org/simple/ pytorch-lightning
```

</details>

### PyTorch Lightning 예제
학습 워크플로우를 정의합니다. 다음은 간단한 토이 예제입니다 ([실제 예제 살펴보기](https://lightning.ai/lightning-ai/studios?view=public&section=featured&query=pytorch+lightning&utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme)):

```python
# main.py
# ! pip install torchvision
import torch, torch.nn as nn, torch.utils.data as data, torchvision as tv, torch.nn.functional as F
import lightning as L

# --------------------------------
# Step 1: LightningModule 정의
# --------------------------------
# LightningModule (nn.Module 서브클래스)은 전체 *시스템*을 정의합니다.
# (예: LLM, diffusion 모델, autoencoder 또는 단순 이미지 분류기).

class LitAutoEncoder(L.LightningModule):
    def __init__(self):
        super().__init__()
        self.encoder = nn.Sequential(nn.Linear(28 * 28, 128), nn.ReLU(), nn.Linear(128, 3))
        self.decoder = nn.Sequential(nn.Linear(3, 128), nn.ReLU(), nn.Linear(128, 28 * 28))

def forward(self, x):
        # lightning에서 forward는 예측/추론 동작을 정의합니다.
        embedding = self.encoder(x)
        return embedding

def training_step(self, batch, batch_idx):
        # training_step은 학습 루프를 정의합니다. forward와는 독립적입니다.
        x, _ = batch
        x = x.view(x.size(0), -1)
        z = self.encoder(x)
        x_hat = self.decoder(z)
        loss = F.mse_loss(x_hat, x)
        self.log("train_loss", loss)
        return loss

def configure_optimizers(self):
        optimizer = torch.optim.Adam(self.parameters(), lr=1e-3)
        return optimizer

# -------------------
# Step 2: 데이터 정의
# -------------------
dataset = tv.datasets.MNIST(".", download=True, transform=tv.transforms.ToTensor())
train, val = data.random_split(dataset, [55000, 5000])

# -------------------
# Step 3: 학습
# -------------------
autoencoder = LitAutoEncoder()
trainer = L.Trainer()
trainer.fit(autoencoder, data.DataLoader(train), data.DataLoader(val))
```

터미널에서 모델 실행

```bash
pip install torchvision
python main.py
```

&nbsp;

# Convert from PyTorch to PyTorch Lightning

PyTorch Lightning은 단지 잘 정리된 PyTorch입니다. Lightning은 PyTorch 코드를 분리하여 과학(모델 로직)을 엔지니어링에서 디커플링합니다.

![PT to PL](https://raw.githubusercontent.com/Lightning-AI/pytorch-lightning/master/docs/source-pytorch/_static/images/general/pl_quick_start_full_compressed.gif)

&nbsp;

----

### Examples
PyTorch Lightning으로 가능한 다양한 유형의 학습을 살펴보세요. 분류, 세그멘테이션, 요약 등 모든 작업을 수행하기 위해 모든 종류의 모델을 Pretrain 및 Finetune 할 수 있습니다.

| Task | Description | Run |
|------|--------------|-----|
| [Hello world](https://lightning.ai/lightning-ai/studios/pytorch-lightning-hello-world?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Pretrain - Hello world 예제 | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/pytorch-lightning-hello-world?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Image classification](https://lightning.ai/lightning-ai/studios/image-classification-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Finetune - 자동차 이미지 분류를 위한 ResNet-34 모델 | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/image-classification-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Image segmentation](https://lightning.ai/lightning-ai/studios/image-segmentation-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Finetune - 이미지 세그멘테이션을 위한 ResNet-50 모델 | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/image-segmentation-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Object detection](https://lightning.ai/lightning-ai/studios/object-detection-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Finetune - 객체 탐지를 위한 Faster R-CNN 모델 | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/object-detection-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Text classification](https://lightning.ai/lightning-ai/studios/text-classification-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Finetune - 텍스트 분류기 (BERT 모델) | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/text-classification-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Text summarization](https://lightning.ai/lightning-ai/studios/text-summarization-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Finetune - 텍스트 요약 (Hugging Face transformer 모델) | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/text-summarization-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Audio generation](https://lightning.ai/lightning-ai/studios/finetune-a-personal-ai-music-generator?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Finetune - 오디오 생성기 (transformer 모델) | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/finetune-a-personal-ai-music-generator?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [LLM finetuning](https://lightning.ai/lightning-ai/studios/finetune-an-llm-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Finetune - LLM (Meta Llama 3.1 8B) | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/finetune-an-llm-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Image generation](https://lightning.ai/lightning-ai/studios/train-a-diffusion-model-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Pretrain - 이미지 생성기 (diffusion 모델) | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/train-a-diffusion-model-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Recommendation system](https://lightning.ai/lightning-ai/studios/recommendation-system-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Train - 추천 시스템 (factorization 및 embedding) | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/recommendation-system-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |
| [Time-series forecasting](https://lightning.ai/lightning-ai/studios/time-series-forecasting-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme) | Train - LSTM을 이용한 시계열 예측 | <a target="_blank" href="https://lightning.ai/lightning-ai/studios/time-series-forecasting-with-pytorch-lightning?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme"><img src="https://pl-bolts-doc-images.s3.us-east-2.amazonaws.com/app-2/studio-badge.svg" alt="Open In Studio"/></a> |

______________________________________________________________________

## Advanced features

Lightning은 대규모의 전문적인 AI 연구를 위해 설계된 [40가지 이상의 고급 기능](https://lightning.ai/docs/pytorch/stable/common/trainer.html?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme#trainer-flags)을 제공합니다.

몇 가지 예시는 다음과 같습니다:

<details>
  <summary>코드 변경 없이 수천 개의 GPU에서 학습</summary>

```python
# 8 GPUs
# 코드 변경이 필요 없습니다.
trainer = Trainer(accelerator="gpu", devices=8)

# 256 GPUs
trainer = Trainer(accelerator="gpu", devices=8, num_nodes=32)
```

</details>

<details>
  <summary>코드 변경 없이 TPU와 같은 다른 가속기에서 학습</summary>

```python
# 코드 변경이 필요 없습니다.
trainer = Trainer(accelerator="tpu", devices=8)
```

</details>

<details>
  <summary>16비트 정밀도 (16-bit precision)</summary>

```python
# 코드 변경이 필요 없습니다.
trainer = Trainer(precision=16)
```

</details>

<details>
  <summary>실험 관리자 (Experiment managers)</summary>

```python
from lightning import loggers

# litlogger
trainer = Trainer(logger=LitLogger())

# tensorboard
trainer = Trainer(logger=TensorBoardLogger("logs/"))

# weights and biases
trainer = Trainer(logger=loggers.WandbLogger())

# comet
trainer = Trainer(logger=loggers.CometLogger())

# mlflow
trainer = Trainer(logger=loggers.MLFlowLogger())

# ... 그리고 수십 가지 더 제공됩니다.
```

</details>

<summary>조기 종료 (Early Stopping)</summary>

```python
es = EarlyStopping(monitor="val_loss")
trainer = Trainer(callbacks=[es])
```

</details>

<details>
  <summary>체크포인팅 (Checkpointing)</summary>

```python
checkpointing = ModelCheckpoint(monitor="val_loss")
trainer = Trainer(callbacks=[checkpointing])
```

</details>

<details>
  <summary>torchscript (JIT)로 내보내기 (운영 환경용)</summary>

```python
# torchscript
autoencoder = LitAutoEncoder()
torch.jit.save(autoencoder.to_torchscript(), "model.pt")
```

</details>

<details>
  <summary>ONNX로 내보내기 (운영 환경용)</summary>

```python
# onnx
with tempfile.NamedTemporaryFile(suffix=".onnx", delete=False) as tmpfile:
    autoencoder = LitAutoEncoder()
    input_sample = torch.randn((1, 64))
    autoencoder.to_onnx(tmpfile.name, input_sample, export_params=True)
    os.path.isfile(tmpfile.name)
```

</details>

______________________________________________________________________

## Advantages over unstructured PyTorch

- 모델이 하드웨어에 구애받지 않게 됩니다 (Hardware agnostic).
- 엔지니어링 코드가 추상화되어 코드를 읽기가 훨씬 명확해집니다.
- 재현성이 높아집니다.
- 까다로운 엔지니어링 부분을 Lightning이 처리하므로 실수가 줄어듭니다.
- 모든 유연성을 유지하면서 (LightningModules는 여전히 PyTorch 모듈입니다) 방대한 양의 상용구 코드(boilerplate)를 제거합니다.
- Lightning은 인기 있는 머신러닝 도구들과 수십 가지의 통합 기능을 제공합니다.
- [모든 새로운 PR에 대해 엄격하게 테스트됩니다](https://github.com/Lightning-AI/lightning/tree/master/tests). 지원되는 모든 PyTorch 및 Python 버전의 조합, 모든 OS, multi GPU 및 TPU까지 테스트합니다.
- 실행 속도 오버헤드가 최소화되었습니다 (순수 PyTorch 대비 에포크당 약 300ms).

______________________________________________________________________

<div align="center">
    <a href="https://lightning.ai/docs/pytorch/stable/?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme">PyTorch Lightning 문서 읽기</a>
</div>

______________________________________________________________________

&nbsp;
&nbsp;

# Lightning Fabric: Expert control

PyTorch 학습 루프와 확장 전략에 대한 전문가 수준의 제어권을 가지고 어떤 장치에서든 어떤 규모로든 실행하세요. 직접 Trainer를 작성할 수도 있습니다.

Fabric은 파운데이션 모델 확장, LLM, diffusion, transformers, 강화 학습, 능동 학습과 같은 매우 복잡한 모델을 위해 설계되었습니다. 규모에 상관없이 적용 가능합니다.

<table>
<tr>
<th>변경 사항</th>
<th>결과 Fabric 코드 (복사해서 사용하세요!)</th>
</tr>
<tr>
<td>
<sub>

```diff
+ import lightning as L
  import torch; import torchvision as tv

dataset = tv.datasets.CIFAR10("data", download=True,
                               train=True,
                               transform=tv.transforms.ToTensor())

+ fabric = L.Fabric()
+ fabric.launch()

model = tv.models.resnet18()
  optimizer = torch.optim.SGD(model.parameters(), lr=0.001)
- device = "cuda" if torch.cuda.is_available() else "cpu"
- model.to(device)
+ model, optimizer = fabric.setup(model, optimizer)

dataloader = torch.utils.data.DataLoader(dataset, batch_size=8)
+ dataloader = fabric.setup_dataloaders(dataloader)

model.train()
  num_epochs = 10
  for epoch in range(num_epochs):
      for batch in dataloader:
          inputs, labels = batch
-         inputs, labels = inputs.to(device), labels.to(device)
          optimizer.zero_grad()
          outputs = model(inputs)
          loss = torch.nn.functional.cross_entropy(outputs, labels)
-         loss.backward()
+         fabric.backward(loss)
          optimizer.step()
          print(loss.data)
```

</sub>
<td>
<sub>

```Python
import lightning as L
import torch; import torchvision as tv

dataset = tv.datasets.CIFAR10("data", download=True,
                              train=True,
                              transform=tv.transforms.ToTensor())

fabric = L.Fabric()
fabric.launch()

model = tv.models.resnet18()
optimizer = torch.optim.SGD(model.parameters(), lr=0.001)
model, optimizer = fabric.setup(model, optimizer)

dataloader = torch.utils.data.DataLoader(dataset, batch_size=8)
dataloader = fabric.setup_dataloaders(dataloader)

model.train()
num_epochs = 10
for epoch in range(num_epochs):
    for batch in dataloader:
        inputs, labels = batch
        optimizer.zero_grad()
        outputs = model(inputs)
        loss = torch.nn.functional.cross_entropy(outputs, labels)
        fabric.backward(loss)
        optimizer.step()
        print(loss.data)
```

</sub>
</td>
</tr>
</table>

## Key features

<details>
  <summary>CPU에서 GPU (Apple Silicon, CUDA, …), TPU, multi-GPU 또는 multi-node 학습으로 쉽게 전환</summary>

```python
# 사용 가능한 하드웨어 사용
# 코드 변경이 필요 없습니다.
fabric = Fabric()

# GPU에서 실행 (CUDA 또는 MPS)
fabric = Fabric(accelerator="gpu")

# 8 GPUs
fabric = Fabric(accelerator="gpu", devices=8)

# 256 GPUs, multi-node
fabric = Fabric(accelerator="gpu", devices=8, num_nodes=32)

# TPU에서 실행
fabric = Fabric(accelerator="tpu")
```

</details>

<details>
  <summary>최신 분산 학습 전략 (DDP, FSDP, DeepSpeed) 및 혼합 정밀도를 즉시 사용</summary>

```python
# 최신 분산 학습 기술 사용
fabric = Fabric(strategy="ddp")
fabric = Fabric(strategy="deepspeed")
fabric = Fabric(strategy="fsdp")

# 정밀도 전환
fabric = Fabric(precision="16-mixed")
fabric = Fabric(precision="64")
```

</details>

<details>
  <summary>모든 장치 로직 상용구(boilerplate)를 자동으로 처리</summary>

```diff
  # 더 이상 필요 없습니다!
- model.to(device)
- batch.to(device)
```

</details>

<details>
  <summary>학습 체크포인팅, 로깅 등을 위한 Fabric 기본 기능을 사용하여 커스텀 Trainer 구축</summary>

```python
import lightning as L

class MyCustomTrainer:
    def __init__(self, accelerator="auto", strategy="auto", devices="auto", precision="32-true"):
        self.fabric = L.Fabric(accelerator=accelerator, strategy=strategy, devices=devices, precision=precision)

def fit(self, model, optimizer, dataloader, max_epochs):
        self.fabric.launch()

model, optimizer = self.fabric.setup(model, optimizer)
        dataloader = self.fabric.setup_dataloaders(dataloader)
        model.train()

for epoch in range(max_epochs):
            for batch in dataloader:
                input, target = batch
                optimizer.zero_grad()
                output = model(input)
                loss = loss_fn(output, target)
                self.fabric.backward(loss)
                optimizer.step()
```

더 광범위한 예제는 [examples](examples/fabric/build_your_own_trainer)에서 확인할 수 있습니다.

</details>

______________________________________________________________________

<div align="center">
    <a href="https://lightning.ai/docs/fabric/stable/?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme">Lightning Fabric 문서 읽기</a>
</div>

______________________________________________________________________

&nbsp;
&nbsp;

## Examples

###### Self-supervised Learning

- [CPC transforms](https://lightning-bolts.readthedocs.io/en/stable/transforms/self_supervised.html#cpc-transforms)
- [Moco v2 transforms](https://lightning-bolts.readthedocs.io/en/stable/transforms/self_supervised.html#moco-v2-transforms)
- [SimCLR transforms](https://lightning-bolts.readthedocs.io/en/stable/transforms/self_supervised.html#simclr-transforms)

###### Convolutional Architectures

- [GPT-2](https://lightning-bolts.readthedocs.io/en/stable/models/convolutional.html#gpt-2)
- [UNet](https://lightning-bolts.readthedocs.io/en/stable/models/convolutional.html#unet)

###### Reinforcement Learning

- [DQN Loss](https://lightning-bolts.readthedocs.io/en/stable/losses.html#dqn-loss)
- [Double DQN Loss](https://lightning-bolts.readthedocs.io/en/stable/losses.html#double-dqn-loss)
- [Per DQN Loss](https://lightning-bolts.readthedocs.io/en/stable/losses.html#per-dqn-loss)

###### GANs

- [Basic GAN](https://lightning-bolts.readthedocs.io/en/stable/models/gans.html#basic-gan)
- [DCGAN](https://lightning-bolts.readthedocs.io/en/stable/models/gans.html#dcgan)

###### Classic ML

- [Logistic Regression](https://lightning-bolts.readthedocs.io/en/stable/models/classic_ml.html#logistic-regression)
- [Linear Regression](https://lightning-bolts.readthedocs.io/en/stable/models/classic_ml.html#linear-regression)

&nbsp;
&nbsp;

## Continuous Integration

Lightning은 여러 CPU, GPU, TPU 및 주요 Python 및 PyTorch 버전에서 엄격하게 테스트됩니다.

###### \*Codecov는 90% 이상이지만 빌드 지연으로 인해 더 낮게 표시될 수 있습니다.

|       System / PyTorch ver.        | 1.13                                                                                                                                                                                                                            | 2.0                                                                                                                                                                                                                             |                                                                                                               2.1                                                                                                               |
| :--------------------------------: | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|
|        Linux py3.9 \[GPUs\]        |  |  | [![Build Status](https://dev.azure.com/Lightning-AI/lightning/_apis/build/status%2Fpytorch-lightning%20%28GPUs%29?branchName=master)](https://dev.azure.com/Lightning-AI/lightning/_build/latest?definitionId=24&branchName=master) |
|  Linux (multiple Python versions)  | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 |                 [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                 |
|   OSX (multiple Python versions)   | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 |                 [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                 |
| Windows (multiple Python versions) | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 | [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                                 |                 [![Test PyTorch](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml/badge.svg)](https://github.com/Lightning-AI/lightning/actions/workflows/ci-tests-pytorch.yml)                 |

</center>
</details>

&nbsp;
&nbsp;

## Community

Lightning 커뮤니티는 다음 구성원들에 의해 유지 관리됩니다.

- 최고의 AI 랩 출신 전문 엔지니어, 연구 과학자, 박사 과정 학생들이 섞인 [10명 이상의 핵심 기여자](https://lightning.ai/docs/pytorch/latest/community/governance.html).
- 800명 이상의 커뮤니티 기여자.

Lightning을 구축하고 수천 명의 연구자들을 위한 상용구 코드를 줄이는 데 도움을 주고 싶으신가요? [첫 기여 방법 알아보기](https://lightning.ai/docs/pytorch/stable/generated/CONTRIBUTING.html?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme)

Lightning은 또한 탄탄한 테스트, 문서화 및 지원이 필요한 [PyTorch 생태계](https://pytorch.org/ecosystem/)의 일부입니다.

### 도움 요청하기

질문이 있으시면 다음 단계를 따라주세요:

1. [문서 읽기](https://lightning.ai/docs?utm_source=ptl_readme&utm_medium=referral&utm_campaign=ptl_readme).
1. [기존 Discussion 검색](https://github.com/Lightning-AI/lightning/discussions) 또는 [새로운 질문 추가](https://github.com/Lightning-AI/lightning/discussions/new).
1. [Discord 채널 참여](https://discord.com/invite/tfXFetEZxv).

pytorch-lightning

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README