This folder contains example scripts for using fastxtend.
imagenette.py allows training on Imagenette and
ImageWoof with most of fastxtend
features as options. It requires Typer.
Run python imagenette.py train -h to get a full list of training
options. And python imagenette.py create -h for FFCV dataset creation
options.
imagenette.yaml is an optional config file for changing the
imagenette.py train defaults. Load the config file by passing
train --config imagenette.yaml to imagenette.py. Passed CLI options
to train will override any config file settings.
To recreate the fastxtend example bencmark on your own system, run:
# warmup & dataset creation
python imagenette.py train --epochs 1
python imagenette.py train --fastai --standard --full-size --epochs 1
# fastai
python imagenette.py train --fastai --standard --full-size
# progressive resizing & fused optimizer with fastai dataloader
python imagenette.py train --fastai
# progressive resizing & fused optimizer with ffcv dataloader
python imagenette.py train