Methods

LightlyTrain supports the following self-supervised learning methods:

• dino (recommended)

  DINO is a popular self-supervised learning method that works well across various datasets, model architectures, and tasks.

• densecl (experimental)

  DenseCL was developed for learning strong local image representations, particularly for detection and segmentation tasks.

• densecldino (experimental)

  DenseCLDINO is a variant of DenseCL that combines DINO and DenseCL for more efficient learning of local image representations.

• simclr

  SimCLR is a classic self-supervised learning method widely used as a baseline for model pre-training.