Why multimodal combines text, image, and audio in Prompt Engineering / GenAI - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load text data for a multimodal model.

Prompt Engineering / GenAI

text_data = load_[1]('data/text_samples.txt')

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

Atext

Baudio

Cvideo

Dimage

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing 'image' or 'audio' instead of 'text'.

The function should load text data, so the correct word is text.

2fill in blank

medium

Complete the code to combine image and audio features for multimodal input.

Prompt Engineering / GenAI

combined_features = concatenate([image_features, [1]_features])

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

Atext

Bvideo

Caudio

Dsensor

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing 'text' or 'video' instead of 'audio'.

We combine image features with audio features, so the blank should be audio.

3fill in blank

hard

Fix the error in the code that processes multimodal inputs by selecting the correct modality.

Prompt Engineering / GenAI

if modality == '[1]':
    process_text(data)
elif modality == 'image':
    process_image(data)
else:
    process_audio(data)

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

Avideo

Baudio

Csensor

Dtext

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'audio' or 'video' instead of 'text'.

The code processes text data when modality is 'text', so the blank must be text.

4fill in blank

hard

Fill both blanks to create a dictionary mapping modalities to their processing functions.

Prompt Engineering / GenAI

processors = {
    'text': [1],
    'image': [2]
}

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

Aprocess_text

Bprocess_audio

Cprocess_image

Dprocess_video

Attempts:

3 left

💡 Hint

Common Mistakes

Swapping functions or using unrelated functions.

The 'text' key maps to process_text and 'image' maps to process_image.

5fill in blank

hard

Fill all three blanks to create a multimodal input pipeline combining text, image, and audio.

Prompt Engineering / GenAI

inputs = {
    'text': [1],
    'image': [2],
    'audio': [3]
}
combined = combine_features([inputs['text'], inputs['image'], inputs['audio']])

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

Atext_features

Bimage_features

Caudio_features

Dvideo_features

Attempts:

3 left

💡 Hint

Common Mistakes

Using video_features or mixing up feature variables.

Text input uses text_features, image uses image_features, and audio uses audio_features.