Training

About Training

Developing a machine learning model usually requires extensive experimentation with different data sets, algorithms, and hyperparameter values. To manage thousands of machine learning model experiments, use the learning feature of ABEJA Platform.

Organize, search and evaluate training jobs using properties, hyperparameters and other data. In addition, learning methods can be selected by selecting cloud or local resources according to the scene, or using templates.

Organize, search and evaluate training jobs and models using the ABEJA Platform console or API.

Guide

• Dataset
• Training handler function
• Configuration file for training
• Local Training
• Template Training (Image Classification)
• Template Training (Object Detection)