Detector BYOD: Training With Your Own Data

If you would like to add your own data to the mix, LEIP Recipes support an easy ingestion of commonly used data formats for classification and detection, such as MS COCO and Pascal. When your data is added to recipes using one of these formats, it verifies that certain conventions are followed and it modifies one of the configuration files to point to the associated components.

Once your data has been provided, the modular nature of LEIP Recipes means that your dataset will be compatible with training future recipes for models of the same type as they are added once your data has been provided. This will give you a simple path for trying out various model sizes and architectures for your application in a reproducible way and will accelerate the path to identifying the best model for your needs.

We provide specific examples for ImageFolder (Classification), COCO, and Pascal VOC (Detection). We will provide a KITTI example in a future release. Please contact Latent AI support if you have an immediate need for KITTI.