Choose a template to use for rendering. Options include 'multimodel' for running multiple models, 'arfi' for ARFI configurations, 'basic', or 'prior'.
Specify the path to the dataset folder containing your input data.
Specify the output folder where the project files will be saved. If the data folder is not in the project-folder, a new data folder will be created and the data will be copied.
Specify the name of the file for managing jobs. By default, this is 'jobs.bat' on Windows and 'jobs.sh' on other platforms.
Provide the file path to a custom template for rendering, if needed.
Select the platform to render jobs for (e.g., Windows, macOS, Linux). If not selected, the current platform will be used by default.
Set the seed for the random selection of priors. This helps ensure reproducibility of the starting set of papers.
Set the seed for the models to control the random number generation used after initialization. Ensures reproducibility.
Set the number of simulation runs to execute. Each run will simulate the same setup with random variations.
The number of records to label before retraining the model. Increasing this value can significantly speed up the simulation.
Skip the generation of dataset wordclouds.
Allow the automatic overwriting of old makita files.
Specify when to stop simulating label actions. Use 'min' to stop when all relevant records are found.
Set -1 to simulate all label actions. Any positive number represents the number of retrainings of the model.
For ARFI templates, set the number of prior records to include. Defaults to 10.
Multi-Model Configurations
Disable the balance strategy for multi-model configurations.
Certain combinations of classifiers and feature extractors do not work in conjunction. Specifying them here will exclude them from the simulations.
Model Configuration
Disable the balance strategy for single-model configurations.