[FEAT] FFCV implementation for dataset

Nobody has claimed this yet.

Hello,

thanks for the library! I am using

if __name__ == "__main__":
# Distill the knowledge from a DINOv3 teacher into a YOLOv8 model
lightly_train.pretrain(
out="out/my_experiment",
data="my_data_dir",
model="ultralytics/yolov8s",
method="distillation",
method_args={
"teacher": "dinov3/vitb16",
},
)

However, using a data directory with the images for SSL can be inefficient. Is there a way to implement this using a .beton (from FFCV library) in order to improve training speed?

Thanks

Start at the lightly_train.pretrain entry point shown in the issue and trace how its data argument is consumed for self-supervised training. Investigate how an FFCV .beton dataset could fit that path; done should mean the feature's scope and supported input behavior are defined, with training-speed improvement assessed.

Written by the indexing model from the issue text.