Quickstart

The platform provides two main ways of usage to meet the needs of different types of users.

1. Graphical User Interface (GUI) Mode

Features

  • Intuitive and user-friendly interface.

  • Supports operations through menus, buttons, and configuration panel, without requiring code.

  • Integrated workflows for common tasks such as data upload, configuration management, and process monitoring.

  • Suitable for non-technical users or first-time users of the platform.

Usage

Step1: Run GUI

  1. Place the EasyCO project folder in the proper directory on your remote server.

  2. Execute the GUI’s Python script run_gui_test.py on your local machine.

Step2: Configure in GUI

  1. Input the server-related information, such as host, port, and username, select the login type, and click the Connect button to connect to the server.

  2. Select a Problem from the list.

  3. Select a Solver from the list.

  4. Set the Parameter of the solver.

  5. Modify Test Parameter or Training Parameter carefully.

Step3: click the button of Start

  1. Click start button to start process.

  2. Get the Log File Output and Summary Output.

  3. Click stop process to terminate the process, or wait until it finishes.

2. Command Line Interface Mode

Features

  • Operates by entering commands in the terminal to run tasks or execute scripts.

  • Provides extensive command options and arguments for customization.

  • Can be seamlessly integrated with existing development toolchains and automation scripts.

  • Suitable for developers, engineers, or advanced users.

Usage

Step1: Basic configuration

Define the fundamental settings—such as problem type, problem size, dataset volume, solver method and so on—in the root configuration file config.yaml.

  • This file controls the overall behavior of the platform.

  • Update these values before starting training or evaluation to match your specific experiment or dataset.

Step2: Solver-specific parameters

Each solver can have its own fine-tuning options.

  • To modify these, open the corresponding YAML file inside the settings directory.

  • For example, you might adjust learning rates, iteration limits, or other algorithm-specific hyperparameters here.

Step3: Running experiments

After configuration is complete, execute the appropriate Python script from the project root:

  • Model testing: python eval.py will load the specified configuration and run the evaluation pipeline.

  • Model training: python train.py will launch the full training workflow and save model checkpoints according to the settings

💡Note: If you prefer not to modify the parameters in the YAML file, you can also set them directly from the command line, for example:

python train.py \
    settings=... \
    mode=... \
    problem=...\
    settings.model.{...}=... \
    settings.model.{...}=... \
    settings.model.{...}=... \

Command-line examples for different methods can be found in the methods introduction section.