IBM SkillsBuild: Build a Simple Image Classifier
Categories
Skills
Project scope
What is the main goal for this project?
The objective of this project is to:
- Create a basic AI model that classifies images and can be used by my business (e.g., handwritten digits or fruit).
Desired outcomes by the end of the project:
- Describe machine learning algorithms and models
- Explain the purpose of IBM Watson Studio
- Describe the key features and benefits of IBM Watson Studio
- Set up a machine learning project in IBM Watson Studio
- Create a Cloud Object Storage resource
- Import a data set into IBM Watson Studio
- Build an AI model using AutoAI in IBM Watson Studio
- Run a prediction experiment for an AI model
- Explain the confusion matrix
- Save a model as a Jupyter Notebook
- Download a notebook in Jupyter Notebook (.ipynb) format
Recommended IBM SkillsBuild course:
The objective of this project is to:
- Create a basic AI model that classifies images and can be used by my business (e.g., handwritten digits or fruit).
Desired outcomes by the end of the project:
- Describe machine learning algorithms and models
- Explain the purpose of IBM Watson Studio
- Describe the key features and benefits of IBM Watson Studio
- Set up a machine learning project in IBM Watson Studio
- Create a Cloud Object Storage resource
- Import a data set into IBM Watson Studio
- Build an AI model using AutoAI in IBM Watson Studio
- Run a prediction experiment for an AI model
- Explain the confusion matrix
- Save a model as a Jupyter Notebook
- Download a notebook in Jupyter Notebook (.ipynb) format
Recommended IBM SkillsBuild course:
What tasks will learners need to complete to achieve the project goal?
By the end of the project, examples of outcomes are:
- Load and preprocess a dataset (e.g., MNIST or CIFAR-10).
- Train and evaluate a simple convolutional neural network.
- Plot and interpret metrics (accuracy, loss curves).
Final deliverables should include:
1. Trained and validated model (e.g., in a notebook or .py script)
2. Codebase with README and model performance results
3. 2–3 page model documentation (architecture, assumptions, limitations)
By the end of the project, examples of outcomes are:
- Load and preprocess a dataset (e.g., MNIST or CIFAR-10).
- Train and evaluate a simple convolutional neural network.
- Plot and interpret metrics (accuracy, loss curves).
Final deliverables should include:
1. Trained and validated model (e.g., in a notebook or .py script)
2. Codebase with README and model performance results
3. 2–3 page model documentation (architecture, assumptions, limitations)