Why AI image generation creates visual content in Prompt Engineering / GenAI - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the library used for AI image generation.

Prompt Engineering / GenAI

import [1]

Drag options to blanks, or click blank then click option'

Atorch

Bnumpy

Ctensorflow

Dmatplotlib

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing numpy or matplotlib which are not primarily for AI model building.

PyTorch is a popular library used for AI image generation models.

2fill in blank

medium

Complete the code to define a simple neural network layer for image generation.

Prompt Engineering / GenAI

layer = torch.nn.Linear([1], 256)

Drag options to blanks, or click blank then click option'

A512

B128

C1024

D784

Attempts:

3 left

💡 Hint

Common Mistakes

Using 512 or 1024 which are common hidden layer sizes but not input size here.

784 is the number of pixels in a 28x28 image, common in image generation tasks.

3fill in blank

hard

Fix the error in the code to generate an image tensor with the right shape.

Prompt Engineering / GenAI

image = torch.randn([1], 3, 64, 64)

Drag options to blanks, or click blank then click option'

A3

B1

C64

D256

Attempts:

3 left

💡 Hint

Common Mistakes

Using 3 which is the number of color channels, not batch size.

The batch size is usually 1 when generating a single image tensor.

4fill in blank

hard

Fill both blanks to create a dictionary that maps image labels to their pixel counts.

Prompt Engineering / GenAI

pixel_counts = {label: [1] for label, pixels in images.items() if len(pixels) [2] 1000}

Drag options to blanks, or click blank then click option'

Alen(pixels)

B>

C<

Dsum(pixels)

Attempts:

3 left

💡 Hint

Common Mistakes

Using sum instead of len, or using < instead of >.

We use len(pixels) to count pixels and filter those with more than 1000 pixels.

5fill in blank

hard

Fill all three blanks to create a dictionary comprehension that selects images with width greater than 50.

Prompt Engineering / GenAI

selected_images = {name: data for name, data in dataset.items() if data['width'] [1] 50 and data['height'] [2] 50 and data['channels'] == [3]

Drag options to blanks, or click blank then click option'

A>

B<

C3

D1

Attempts:

3 left

💡 Hint

Common Mistakes

Using < instead of > or wrong channel number.

We check if width and height are greater than 50 and channels equal 3 for color images.

Practice

1. Why does AI image generation create pictures from text descriptions?

easy

A. To calculate numbers faster

B. To write long stories automatically

C. To turn ideas into visual images that are easy to understand

D. To translate languages word by word

Solution

1. Step 1: Understand the purpose of AI image generation

   AI image generation uses text input to create pictures that show ideas visually.

2. Step 2: Match the purpose with the options

   Only To turn ideas into visual images that are easy to understand explains that AI turns ideas into images for easier understanding.

3. Final Answer:

   To turn ideas into visual images that are easy to understand -> Option C

4. Quick Check:

   AI image generation = visual idea creation [OK]

Hint: AI image generation = text to pictures [OK]

Common Mistakes:

• Confusing image generation with text writing
• Thinking AI only translates languages
• Assuming AI calculates numbers

2. Which of these is the correct way to give a prompt for AI image generation?

easy

A. Draw a cat sitting on a red chair

B. Calculate 5 plus 3

C. Translate hello to Spanish

D. Write a poem about trees

Solution

1. Step 1: Identify the prompt type for image generation

   AI image generation needs a description of what to draw, like an object and setting.

2. Step 2: Check which option describes a visual scene

   Draw a cat sitting on a red chair describes a cat on a red chair, which is a clear visual prompt.

3. Final Answer:

   <code>Draw a cat sitting on a red chair</code> -> Option A

4. Quick Check:

   Visual prompt = Draw a cat sitting on a red chair [OK]

Hint: Prompts for images describe scenes or objects [OK]

Common Mistakes:

• Using commands for math or translation instead of images
• Writing text tasks instead of visual descriptions
• Confusing image prompts with text generation

3. What will the AI most likely generate from this prompt? 'A sunny beach with palm trees and blue water'

medium

A. A graph showing temperature changes at the beach

B. A picture showing a sunny beach with palm trees and blue water

C. A list of beach locations worldwide

D. A text story about a beach vacation

Solution

1. Step 1: Understand the prompt content

   The prompt describes a visual scene: sunny beach, palm trees, blue water.

2. Step 2: Match the prompt to the AI output type

   AI image generation creates pictures, so it will produce an image matching the description.

3. Final Answer:

   A picture showing a sunny beach with palm trees and blue water -> Option B

4. Quick Check:

   Visual prompt = visual image output [OK]

Hint: Visual description prompt = image output [OK]

Common Mistakes:

• Expecting text or lists instead of images
• Confusing AI image generation with text generation
• Thinking AI outputs graphs from text

4. An AI image generator is given the prompt 'A red apple on a table' but outputs a blue apple. What is the likely cause?

medium

A. The prompt was too short and unclear

B. The AI cannot generate images of apples

C. The AI only creates black and white images

D. The AI misunderstood the color word in the prompt

Solution

1. Step 1: Analyze the prompt and output mismatch

   The prompt says 'red apple' but output shows a blue apple, so color was misunderstood.

2. Step 2: Check other options for correctness

   The AI can generate apples and colors; prompt length is sufficient; AI can create color images.

3. Final Answer:

   The AI misunderstood the color word in the prompt -> Option D

4. Quick Check:

   Color mismatch = misunderstanding prompt [OK]

Hint: Color errors usually mean prompt misunderstanding [OK]

Common Mistakes:

• Blaming AI for inability to create objects it can make
• Assuming prompt length is always the problem
• Thinking AI only makes black and white images

5. You want an AI to create a detailed image of a futuristic city at night with neon lights. Which prompt will most likely produce the best image?

hard

A. 'A futuristic city at night with bright neon lights and flying cars'

B. 'City with buildings'

C. 'Night scene'

D. 'A city during the day with trees'

Solution

1. Step 1: Compare prompt details

   'A futuristic city at night with bright neon lights and flying cars' has the most detailed description including time, style, lighting, and objects.

2. Step 2: Understand how detail affects AI image quality

   More details in the prompt help AI create accurate and rich images matching the idea.

3. Final Answer:

   'A futuristic city at night with bright neon lights and flying cars' -> Option A

4. Quick Check:

   Detailed prompt = better image [OK]

Hint: More details in prompt = better images [OK]

Common Mistakes:

• Using vague or short prompts
• Ignoring important scene details like time or lighting
• Choosing unrelated scene descriptions