Bùi Nguyễn Kim Hải

Bùi Nguyễn Kim Hải

he/him

BSc CS @ELTE | Backend Developer Intern @Butopêa | Undergraduate Researcher

[5-Day Gen AI Intensive Course] Day 1: Prompting

2 minute read

Published:

What are the interesting things in the first day of 5-Day Gen AI Intensive Course?

Disclaimer: This write-up’s content is mainly about what I feel interesting in the Day 1 of 5-Day Gen AI Intensive Course, so it will not contain every that is mentioned in the course. To have more information, please access the course’s contents:

First of all, Import & Config

%pip install -U -q "google-generativeai>=0.8.3"

import google.generativeai as genai
from IPython.display import HTML, Markdown, display

GOOGLE_API_KEY = 'YOUR_SECRET_KEY'
genai.configure(api_key=GOOGLE_API_KEY)

Chat History

I have worked with Gemini API since August of 2024, but up to now, I don’t know that we can send the chat history for the LLM. Let’s see how they set it up.

flash = genai.GenerativeModel('gemini-1.5-flash')
chat = flash.start_chat(history=[])

response = chat.send_message('Hello! My name is Zlork.')
print(response.text)
# Hello Zlork! It's nice to meet you. 😊 What can I do for you today?

response = chat.send_message('Can you tell something interesting about dinosaurs?')
print(response.text)
# ... Response

response = chat.send_message('Do you remember what my name is?')
print(response.text)
# Yes, I remember! You said your name is Zlork. 😊 ...

Enum Mode

This is absolutely new to me! Let’s see how to write.

import enum

class Sentiment(enum.Enum):
    POSITIVE = "positive"
    NEUTRAL = "neutral"
    NEGATIVE = "negative"


model = genai.GenerativeModel(
    'gemini-1.5-flash-001',
    generation_config=genai.GenerationConfig(
        response_mime_type="text/x.enum",
        response_schema=Sentiment
    ))

response = model.generate_content(zero_shot_prompt)
print(response.text)
# positive

I think this will be the best for ones who want to try classification tasks.

JSON mode

I know about JSON mode, but I’ve never think it could be this way. All of my past experience is nothing.

import typing_extensions as typing

class PizzaOrder(typing.TypedDict):
    size: str
    ingredients: list[str]
    type: str

model = genai.GenerativeModel(
    'gemini-1.5-pro-002',
    generation_config=genai.GenerationConfig(
        temperature=0.1,
        response_mime_type="application/json",
        response_schema=PizzaOrder,
    ))

response = model.generate_content("Can I have a large dessert pizza with apple and chocolate")
print(response.text)
# {"ingredients": ["apple", "chocolate"], "size": "large", "type": "dessert"}

Or even combine them together!

import enum
from typing_extensions import TypedDict

class Grade(enum.Enum):
    A_PLUS = "a+"
    A = "a"
    B = "b"
    C = "c"
    D = "d"
    F = "f"

class Recipe(TypedDict):
    recipe_name: str
    grade: Grade

model = genai.GenerativeModel("gemini-1.5-pro-latest")

result = model.generate_content(
    "List about 10 cookie recipes, grade them based on popularity",
    generation_config=genai.GenerationConfig(
        response_mime_type="application/json", response_schema=list[Recipe]
    ),
)
print(result)  # [{"grade": "a+", "recipe_name": "Chocolate Chip Cookies"}, ...]