Topic: “Transfer Learning with Tensorflow Hub”
Speaker: Weiguang Guan, SHARCNET
Transfer learning is an important learning technique that can be used to leverage pre-trained models on new similar problems. More and more successful deep learning applications are developed on top of previously trained models rather than from scratch. Tensorflow Hub is a fast growing repository that hosts a large collection of pre-trained models for reuse. In this tutorial, we will introduce what you need to know in order to use this useful technique. The entire process of transfer learning is illustrated through a case study, in which we will discuss (1) how to use a pre-trained model as a feature extractor and (2) how to fine-tune the added layers with the base model together.
Zoom account is required. Webinar registration is required. VERY IMPORTANT: During registration, you have to provide the email address which is associated with your Zoom account, otherwise you won’t be able to register and/or attend the webinar! Need help attending a webinar? See the SHARCNET Help Wiki.