we’ll fine-tune a VGG mannequin with PyTorch on the Caltech-101 dataset whereas showcasing how leveraging a pre-trained mannequin can enhance efficiency on new duties
Switch studying is a way the place a pre-trained mannequin is repurposed for a brand new process. For instance, say you already know learn how to prepare dinner Italian meals properly and need to department out into Chinese language cooking. You don’t have to relearn learn how to cube an onion, so the abilities you have already got could be transferred to Chinese language delicacies. You’ll want just a few new strategies, however you’ve already acquired a stable baseline to start out with.
That’s transition studying in a nutshell.
With switch studying, you’re taking the pre-trained community (Italian cooking) and fine-tune it to be taught Chinese language dishes. You leverage the data it gained from earlier duties to assist it be taught the brand new process. This protects a ton of time and computational sources in comparison with coaching a community from scratch.
PyTorch is the toolkit we’ll use to implement this. Consider it as your kitchen, the place you’ve gotten all of the instruments and elements to prepare dinner up your machine-learning fashions. PyTorch is tremendous well-liked as a result of it’s beginner-friendly, versatile, and has nice group assist. Plus, it’s acquired tons of helpful options to make constructing and coaching neural networks a breeze.
