Welcome to our Support Center
< All Topics

Introduction: iERP.ai Studio application basics

Introduction to iERP.ai Studio application


iERP.ai Studio application is a standalone product for companies of all sizes allowing to train and deploy Artificial Intelligence (AI) algorithms safely, quickly and without any prior knowledge of AI. The application has the following key concepts:
iERP.ai Studio application (GUI mode)
In this mode, the application is installed and started with the graphical user interface. This allows the consultant to run all functions of the application very easily.
iERP.ai Studio application (CLI mode)
This command-line mode is used when the application is installed on the customer server as a service. The typical workflow is that a consultant will use GUI mode to setup customer project, select and define parameters for the algorithm, select data source and verify data quality. Once project is completely prepared, the consultant will install Studio in CLI mode on the customer server and let it run automatically for months without any intervention.
Studio application – Projects, Trainer & Forecaster
Studio application has 4 key sections:
  • Projects – iERP.ai Studio is working with the concept of projects. Each time an algorithm is implemented for a customer, a new project needs to be created. All files for a specific project are stored in a project folder. One algorithm only can be selected in single project. Backups and any other maintenance activities could be simply performed by copying the project folder with all files stored within it.
  • Trainer – A set of 4 screens used for Algorithm training. Output from the 4th screen is a fully trained model, based on customer-specific data. The Trainer section is used for data upload, data validation, data review, and algorithm machine learning (ML) training. Some algorithms (Discount recommendation) are distributed with pre-trained models therefore if they are selected, only the forecaster section is available.
  • Forecaster: – A set of 4 screens using a trained model from the Trainer to estimate expected output based on an algorithm. The Forecaster section is used for most recent data load, data validation, data review, business intelligence, forecast generation/export/review and setup of additional features as Like functionality to market Intelligence functionality.
  • Production: – Use the Production screen to install API as a service, review algorithm training history and forecasting history, and review logging.
Studio application – Help files
Studio application has a content-sensitive help file structure which is displaying information about individual screens and fields located on the screens.

Next Installation: Studio application hardware and software requirements
Table of Contents