Transfer Learning in fast.ai - How does the magic work?

Deep Learning
17

@TomB You’re right, running a PyTorch model with a fast.ai Learning was much easier. Just did that, and I got exactly the same results, as when using the fast.ai model (i.e. the resnet50 model that fast.ai uses seems to be same that PyTorch uses).

Here the code I used:

import torchvision
learn = cnn_learner(data, torchvision.models.resnet50, metrics=[accuracy], true_wd=False)
learn.fit(3)

So the magic happens definitely somewhere in the data pipeline… Very strange…