from flush import FewShotPromptTemplate

Let’s first look at a very simple example of how a FewShotPromptTemplate can be used to format a prompt:

​
Creating Examples

To get started, we can create a list of few-shot examples. Each example is a tuple that contains the input first and the formatted output next.

examples = [
    ('''Write a compressed perfect image description with adjectives and nouns 
    of a Female cyborg walking in a winter landscape in Norway:''', 
    '''Gorgeous female cyborg, shimmering, sci-fi, armor, strides, snow, tree, 
    banks, frosted, ice, gleaming, metal, blue, optics, robotic, movements, still, 
    pristine, beauty, Nordic, vista'''),

    ('''Write a compressed perfect image description with adjectives and nouns 
    of a Female cyborg walking in a winter landscape in Norway:''', 
    '''Beautiful female cyborg, powerful sci-fi armor, snow, trees, banks, 
    frosted, ice, gleaming, metal, white, high tech, robotic, still, pristine, 
    majestic, Nordic, epic'''),

    ('''Write a compressed perfect image description with adjectives and nouns 
    of a Female cyborg walking in a winter landscape in Norway:''', 
    '''Magnificent female cyborg, dazzling futuristic armor, snow-covered, trees, 
    banks, iced, chilly, shining, metallic, electric-blue, lenses, precise, 
    mechanical, motion, serene, picturesque, glorious, Nordic''')
]

​
Defining the Main Template

Now, we define the input that we will pass along with the few shot examples. We use Python’s f-string as the template format for prompt templates.

main_template = '''Write a compressed perfect image description with 
                   adjectives and nouns of a Female cyborg walking in a 
                   {season} landscape in Norway:?'''

This can then be passed into model input.

​
Feeding the Examples and Main Template to FewShotPromptTemplate

Finally, create a FewShotPromptTemplate object. This boject takes in the few-shot examples, as well as the main template.

introduction_text = "You are to construct prompts for AI image generation models."

prompt_template = FewShotPromptTemplate(
    examples=examples,
    main_template=main_template,
    introduction=introduction_text
)

Note that we also pass in introduction, an optional parameter simply used to clarify the central theme of the examples if needed.

​
Formatting the Template

Now, we use the format method to create a prompt with new content as shown below.

formatted_prompt = prompt_template.format(season="summer")
print(formatted_prompt)

This will give us the following output:

You are to construct prompts for AI image generation models.

Write a compressed perfect image description with adjectives and nouns 
of a Female cyborg walking in a winter landscape in Norway:
Gorgeous female cyborg, shimmering, sci-fi, armor, strides, snow, tree, 
banks, frosted, ice, gleaming, metal, blue, optics, robotic, movements, still, 
pristine, beauty, Nordic, vista

Write a compressed perfect image description with adjectives and nouns 
of a Female cyborg walking in a winter landscape in Norway:
Beautiful female cyborg, powerful sci-fi armor, snow, trees, banks, 
frosted, ice, gleaming, metal, white, high tech, robotic, still, pristine, 
majestic, Nordic, epic

Write a compressed perfect image description with adjectives and nouns 
of a Female cyborg walking in a winter landscape in Norway:
Magnificent female cyborg, dazzling futuristic armor, snow-covered, trees, 
banks, iced, chilly, shining, metallic, electric-blue, lenses, precise, 
mechanical, motion, serene, picturesque, glorious, Nordic

Write a compressed perfect image description with adjectives and nouns 
of a Female cyborg walking in a summer landscape in Norway:

from flush import FewShotPromptTemplate

Let’s first look at a very simple example of how a FewShotPromptTemplate can be used to format a prompt:

​
Creating Examples

To get started, we can create a list of few-shot examples. Each example is a tuple that contains the input first and the formatted output next.

examples = [
    ('''Write a compressed perfect image description with adjectives and nouns 
    of a Female cyborg walking in a winter landscape in Norway:''', 
    '''Gorgeous female cyborg, shimmering, sci-fi, armor, strides, snow, tree, 
    banks, frosted, ice, gleaming, metal, blue, optics, robotic, movements, still, 
    pristine, beauty, Nordic, vista'''),

    ('''Write a compressed perfect image description with adjectives and nouns 
    of a Female cyborg walking in a winter landscape in Norway:''', 
    '''Beautiful female cyborg, powerful sci-fi armor, snow, trees, banks, 
    frosted, ice, gleaming, metal, white, high tech, robotic, still, pristine, 
    majestic, Nordic, epic'''),

    ('''Write a compressed perfect image description with adjectives and nouns 
    of a Female cyborg walking in a winter landscape in Norway:''', 
    '''Magnificent female cyborg, dazzling futuristic armor, snow-covered, trees, 
    banks, iced, chilly, shining, metallic, electric-blue, lenses, precise, 
    mechanical, motion, serene, picturesque, glorious, Nordic''')
]

​
Defining the Main Template

Now, we define the input that we will pass along with the few shot examples. We use Python’s f-string as the template format for prompt templates.

main_template = '''Write a compressed perfect image description with 
                   adjectives and nouns of a Female cyborg walking in a 
                   {season} landscape in Norway:?'''

This can then be passed into model input.

​
Feeding the Examples and Main Template to FewShotPromptTemplate

Finally, create a FewShotPromptTemplate object. This boject takes in the few-shot examples, as well as the main template.

introduction_text = "You are to construct prompts for AI image generation models."

prompt_template = FewShotPromptTemplate(
    examples=examples,
    main_template=main_template,
    introduction=introduction_text
)

Note that we also pass in introduction, an optional parameter simply used to clarify the central theme of the examples if needed.

​
Formatting the Template

Now, we use the format method to create a prompt with new content as shown below.

formatted_prompt = prompt_template.format(season="summer")
print(formatted_prompt)

This will give us the following output:

You are to construct prompts for AI image generation models.

Write a compressed perfect image description with adjectives and nouns 
of a Female cyborg walking in a winter landscape in Norway:
Gorgeous female cyborg, shimmering, sci-fi, armor, strides, snow, tree, 
banks, frosted, ice, gleaming, metal, blue, optics, robotic, movements, still, 
pristine, beauty, Nordic, vista

Write a compressed perfect image description with adjectives and nouns 
of a Female cyborg walking in a winter landscape in Norway:
Beautiful female cyborg, powerful sci-fi armor, snow, trees, banks, 
frosted, ice, gleaming, metal, white, high tech, robotic, still, pristine, 
majestic, Nordic, epic

Write a compressed perfect image description with adjectives and nouns 
of a Female cyborg walking in a winter landscape in Norway:
Magnificent female cyborg, dazzling futuristic armor, snow-covered, trees, 
banks, iced, chilly, shining, metallic, electric-blue, lenses, precise, 
mechanical, motion, serene, picturesque, glorious, Nordic

Write a compressed perfect image description with adjectives and nouns 
of a Female cyborg walking in a summer landscape in Norway: