Deploying with Docker

To deploy SwanLab Community Edition privately, follow the installation steps below.

Prerequisites

Before installing SwanLab, ensure your machine meets the following minimum system requirements:

• CPU >= 2 cores
• Memory >= 4GB
• Storage space >= 20GB

SwanLab Community Edition requires Docker Compose for installation and deployment. Refer to the table below to select the correct Docker and Compose versions based on your operating system.

If Docker is already installed, skip this step.

Operating System	Software	Notes
macOS 10.14 or later	Docker Desktop	Configure the Docker VM to use at least 2 vCPUs and 8 GB of initial memory. Otherwise, installation may fail. For more details, refer to the Docker Desktop for Mac Installation Guide.
Windows (with WSL 2 enabled)	Docker Desktop	We recommend storing source code and other Linux container-bound data in the Linux filesystem rather than the Windows filesystem. For more details, refer to the Docker Desktop for Windows with WSL 2 Backend Installation Guide.
Linux	Docker 19.03 or later, Docker Compose 1.28 or later	For installation instructions, refer to the Docker Installation Guide and Docker Compose Installation Guide.

If Docker is not installed, you can run the provided installation script.

1. Clone the Repository

Clone the self-hosted repository using Git:

git clone https://github.com/SwanHubX/self-hosted.git
cd self-hosted

2. One-Click Installation

If you are using Windows, please enable WSL2 for installation.

The default installation script is located at docker/install.sh. Execute it to install all required containers and perform initial configurations.

cd ./docker
./install.sh

The default script link uses a mirror source in China, so the download speed in China is very fast!

If you prefer to use DockerHub as the image registry, run the following script:

./install-dockerhub.sh

3. Activate the Primary Account

SwanLab Community Edition uses port 8000 by default. If using the default configuration, access it at: http://localhost:8000.

If the community edition is deployed on a different port, open Docker Desktop, locate the port mapping for the traefik container (e.g., 64703:80), and access http://localhost:64703.

Now, activate your primary account. Activation requires 1 License, which can be obtained for free from the SwanLab Official Website for personal use.

Once you have the License, return to the activation page, enter your username, password, confirm password, and License, then click "Activate" to complete the setup.

4. Start Your First Experiment

Log in using the Python SDK:

swanlab login --host <IP Address>

If you have previously logged in and wish to re-login, use:
swanlab login --host <IP Address> --relogin.

Press Enter, enter your API Key, and complete the login. Your SwanLab experiments will now be uploaded to your private SwanLab instance by default.

---

Test Script:

import swanlab
import random

# Create a SwanLab project
swanlab.init(
    # Set project name
    project="my-awesome-project",
    
    # Set hyperparameters
    config={
        "learning_rate": 0.02,
        "architecture": "CNN",
        "dataset": "CIFAR-100",
        "epochs": 10
    }
)

# Simulate a training session
epochs = 10
offset = random.random() / 5
for epoch in range(2, epochs):
  acc = 1 - 2 ** -epoch - random.random() / epoch - offset
  loss = 2 ** -epoch + random.random() / epoch + offset

  # Log training metrics
  swanlab.log({"acc": acc, "loss": loss})

# [Optional] Complete the training, necessary in notebook environments
swanlab.finish()

View the experiment on the web after running: