GenAI applications (text, image, code, audio) - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Which GenAI application is best suited for generating realistic human speech?

Imagine you want to create a virtual assistant that talks like a real person. Which GenAI application would you choose?

A An audio generation model that synthesizes speech.

B An image generation model that creates photos.

C A text generation model that writes stories.

D A code generation model that writes software scripts.

💡 Hint

Think about which type of data sounds like human voice.

Audio generation models create realistic speech sounds, making them ideal for virtual assistants that talk.

❓ Predict Output

intermediate

What is the output of this text generation code snippet?

Given the following code using a GenAI text model, what will be the output?

Prompt Engineering / GenAI

prompt = "Hello, how are you?"
response = genai_text_model.generate(prompt)
print(response)

A SyntaxError: missing parentheses in call to 'print'

B "Hello, how are you? I am fine, thank you!"

C "<image>"

D "def hello(): pass"

💡 Hint

The model continues the text prompt with a natural response.

The text generation model takes the prompt and produces a natural language continuation.

❓ Model Choice

advanced

Which GenAI model type is most appropriate for creating detailed artwork from a text description?

You want to generate a detailed painting from a sentence like "A sunset over mountains." Which model type should you use?

A Image generation model

B Text generation model

C Audio generation model

D Code generation model

💡 Hint

Think about which model creates pictures.

Image generation models convert text descriptions into images, perfect for artwork creation.

❓ Metrics

advanced

Which metric best evaluates the quality of generated code by a GenAI model?

You have a GenAI model that writes code. Which metric helps measure how well the generated code works?

A Accuracy of image classification

B Word error rate in speech recognition

C BLEU score comparing generated code to reference code

D Mean squared error of audio signals

💡 Hint

Think about comparing generated text to a known correct version.

BLEU score measures similarity between generated and reference code, useful for code generation quality.

🔧 Debug

expert

What error does this audio generation code produce?

Consider this code snippet for generating audio with a GenAI model. What error will it raise?

Prompt Engineering / GenAI

audio = genai_audio_model.generate(text_input='Hello world')
print(audio.shape)

A No error, prints the shape of the audio tensor

B TypeError: unsupported operand type(s) for +: 'int' and 'str'

C ValueError: invalid literal for int() with base 10

D AttributeError: 'NoneType' object has no attribute 'shape'

💡 Hint

Check if the generate method returns a valid audio object.

If the generate method returns None (e.g., due to missing model setup), accessing .shape causes AttributeError.