From the course: Computer Vision for Data Scientists

Unlock the full course today

Join today to access over 24,900 courses taught by industry experts.

Introduction to transfer learning

Introduction to transfer learning

From the course: Computer Vision for Data Scientists

Start my 1-month free trial Buy for my team

Introduction to transfer learning

“

- [Instructor] When creating a new deep learning model, it could be challenging to gather enough labeled data, computational resources, and time. Fortunately, transfer Learning provides an efficient solution to this issue. This technique involves pre-training a model for one task and then fine tuning it for another related task. Transfer learning is an important technique in deep learning as it saves time and computational resources that would otherwise be spent training large models from scratch. The initial layers of a pre-trained model are trained to recognize low level features like edges and textures, which are then transferred to the new task. This is reducing the amount of training required on large amounts of labeled data. Transfer learning often leads to better generalization because the model has already learned useful feature representations from the pre-training task. Transfer learning offers several benefits for…

Contents