ML project structure in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Identify the main purpose of the 'data' folder in an ML project

In a typical machine learning project, what is the main purpose of the data folder?

A To store raw and processed datasets used for training and testing models

B To save the final trained model files and checkpoints

C To keep scripts for data visualization and plotting results

D To hold configuration files for project settings

💡 Hint

Think about where you keep the information your model learns from.

The data folder is where all raw and processed datasets are stored. This includes training, validation, and test data. It is essential for the model to access this data during training and evaluation.

❓ Model Choice

intermediate

Choosing the right folder for model code

You have written Python code that defines your machine learning model architecture and training functions. Where should you place this code in a well-organized ML project?

A Inside a folder named <code>models</code> or <code>src/models</code>

B In the <code>data</code> folder alongside datasets

C In the root folder without any subfolder

D Inside the <code>notebooks</code> folder

💡 Hint

Think about where code related to model design logically belongs.

The models folder or a similar src/models path is the best place to keep model architecture and training code. This keeps code organized and separate from data and experiments.

❓ Metrics

advanced

Understanding evaluation metrics placement

Where in an ML project structure should you place code that calculates and logs evaluation metrics like accuracy or loss during model training?

A Directly inside the <code>data</code> folder

B Inside the <code>evaluation</code> or <code>metrics</code> folder or module

C In the <code>models</code> folder with the model architecture code

D In the <code>config</code> folder

💡 Hint

Think about where you keep code related to checking how well your model performs.

Evaluation metric code is best kept in a dedicated evaluation or metrics folder or module. This separates concerns and makes it easier to update or add new metrics.

🔧 Debug

advanced

Debugging misplaced files in ML project

You find that your ML project has training scripts inside the data folder and dataset files inside the models folder. What is the main problem caused by this structure?

A Training scripts will run faster due to proximity to data files

B The model will not train because data files are in the wrong folder

C It causes confusion and makes the project hard to maintain and understand

D There is no problem; folders can contain any files

💡 Hint

Think about how clear and easy it is to find things in a project.

Misplacing files causes confusion and makes the project difficult to maintain. Clear separation of data, code, and models helps teamwork and debugging.

❓ Hyperparameter

expert

Best practice for managing hyperparameters in ML projects

In a well-structured ML project, where should you store hyperparameters like learning rate, batch size, and number of epochs to ensure easy tuning and reproducibility?

A Hardcoded directly inside the training script code

B In the <code>models</code> folder alongside model architecture code

C Inside the <code>data</code> folder with datasets

D In a separate <code>config</code> file or folder, such as YAML or JSON files

💡 Hint

Think about how to change settings without editing code.

Storing hyperparameters in separate config files like YAML or JSON allows easy tuning without changing code. It also helps track experiments and reproduce results.