Key models overview (GPT, DALL-E, Stable Diffusion) in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

What is the primary function of GPT models?

GPT models are designed mainly to:

A Enhance audio signals for speech clarity

B Generate and understand human-like text

C Create detailed images from text descriptions

D Predict stock market trends using numerical data

💡 Hint

Think about what GPT stands for and what it is famous for.

GPT (Generative Pre-trained Transformer) models specialize in generating and understanding text, making them powerful for chatbots, writing, and language tasks.

❓ Predict Output

intermediate

Output of a Stable Diffusion image generation step

Given a text prompt, what type of output does Stable Diffusion produce?

A A sound clip related to the prompt

B A text summary describing the prompt

C A high-resolution image matching the prompt

D A 3D model file of the described object

💡 Hint

Stable Diffusion is known for creating visuals from words.

Stable Diffusion generates images from text prompts by gradually refining noise into a detailed picture.

❓ Model Choice

advanced

Choosing the right model for text-to-image generation

You want to create an artistic image from a text description. Which model is best suited?

A Stable Diffusion

B GPT

C DALL-E

D BERT

💡 Hint

Consider which models specialize in image creation from text.

Stable Diffusion is a popular open-source model for generating images from text prompts, known for artistic and detailed outputs.

❓ Metrics

advanced

Evaluating GPT model text generation quality

Which metric is commonly used to measure how well GPT models generate coherent and relevant text?

A Fréchet Inception Distance

B Inception Score

C Mean Squared Error

D BLEU score

💡 Hint

Think about metrics used in natural language processing.

BLEU score measures how closely generated text matches reference text, commonly used for language models like GPT.

🔧 Debug

expert

Why might a DALL-E model fail to generate an image from a prompt?

Given a valid text prompt, which of the following is the most likely cause for DALL-E not producing an image?

A The prompt contains unsupported special characters causing tokenization errors

B The model is trained only on audio data, so it cannot generate images

C The prompt is too short, so the model refuses to run

D The model requires GPU but is running on CPU, which is unsupported

💡 Hint

Consider how text inputs are processed before image generation.

DALL-E tokenizes text prompts; unsupported characters can cause tokenization errors preventing image generation.