Embedded Files
As we briefly mentioned in Module 3, a transfer learning model is a machine learning model that uses a model weights which were previously trained for an original task to be used on a new, related task. In contrast to creating a completely new model, a transfer learning model has imported weights from its previous training and adapts these weights to better fit the new task at hand. In general, using a transfer learning model requires less training data and computational power. This activity will introduce you to a few of the benefits of using a transfer learning model as compared to a self-written and compiled machine learning model.
This Transfer Learning Models Activity will help you:
This Transfer Learning Models Activity will help you:
Understand the advantages of using transfer learning models
Experience how to import a pre-built machine learning model into Google Colab
Use the pre-built YOLOv8 model to predict objects within an image
Evaluate the performance of a transfer learning model via its predictions
This Transfer Learning Models Activity will help you:
This Transfer Learning Models Activity will help you:
Understand the advantages of using transfer learning models
Experience how to import a pre-built machine learning model into Google Colab
Use the pre-built YOLOv8 model to predict objects within an image
Evaluate the performance of a transfer learning model via its predictions
Required Materials/Resources:
Required Materials/Resources:
Access to a computer and large screen (if you want share with others)
A Google account to access Google Colab
Keywords:
Keywords:
YOLOv8 Model - A transfer learning model released by the company Ultralytics in January of 2023. This model is optimized for real-time object detection applications. There is a YOLOv8 model for an array of sub-tasks such as detection, segmentation, pose or keypoint detection, oriented detection, and classification.
Command Shell -
Background:
Background:
Transfer learning is a powerful tool in machine learning applications as it allows for the fast and easy creation of general machine learning models that are quick to train and require less training time since the model is mostly pre-trained already. This is especially useful when you don't have a lot of training data available to you or the equipment to train a machine learning model that requires a lot of data and processing power to train.
In this activity we will be looking at how transfer learning works in machine learning, how to implement and transfer a model and how learning this skill can allow you to create different kinds of models not originally available to you due to lacking the amount of datasets or hardware required to train a complex model.
Activity Layout & Instructions
Activity Layout & Instructions
- Preparation
Upload an image of an animal of your own to the Module 4 folder of the Image Processing Image Dataset folder we downloaded in the Module 1 Introduction Activity and name it as: yoloimage.jpg or yoloimage2.jpg.
- Activity
Run each of the sections within the Transfer Learning Model - Colab Notebook and observe from your uploaded image.
Downloadable Content & Links
Downloadable Content & Links
Further Learning Resources
Further Learning Resources
Page updated
Report abuse