mapper/models/dinov2/configs/train/vitl16_short.yaml

# this corresponds to the default config
train:
  dataset_path: ImageNet:split=TRAIN
  batch_size_per_gpu: 64
student:
  block_chunks: 4