Pytorch models trained on not-so-large datasets

When looking at the pretrained classification models from pytorch, they all seem to be trained on very large datasets such as ImageNet, etc.

Is it possible for me to find some pretrainend models somewhere that are compatible with pytorch and are trained with datasets that aren’t extremely large (so that I can use them on my laptop)? Preferably classification models. I am planning to use these models to run some compression experiments for which I need part of the training data, as well as a test or validation set.

Maybe I didn’t understand the question well, wouldn’t it be more convenient to use Google Colab (you can upload the data to G Drive).
Alternatively, locally, you can use