Usar o Gemini para identificar momentos importantes em vídeos do YouTube

Este exemplo de código demonstra como usar o Gemini para identificar os principais momentos em vídeos do YouTube. Ele usa um URL de vídeo do YouTube como entrada e retorna uma lista de momentos importantes com os respectivos carimbos de data/hora.

Exemplo de código

Python

Antes de testar esse exemplo, siga as instruções de configuração para Python no Guia de início rápido da Vertex AI sobre como usar bibliotecas de cliente. Para mais informações, consulte a documentação de referência da API Vertex AI para Python.

Para autenticar na Vertex AI, configure o Application Default Credentials. Para mais informações, consulte Configurar a autenticação para um ambiente de desenvolvimento local.

import vertexai
from vertexai.generative_models import GenerativeModel, Part

# TODO (developer): update project id
vertexai.init(project=PROJECT_ID, location="us-central1")

model = GenerativeModel("gemini-1.5-flash-002")

contents = [
    # Text prompt
    "Identify the key moments of this video.",
    # YouTube video of Paris 2024 Olympics
    Part.from_uri("https://www.youtube.com/watch?v=6F5gZWcpNU4", "video/mp4"),
]

response = model.generate_content(contents)
print(response.text)
# Example response
#    This video is a fast-paced, exciting montage of athletes competing in and celebrating their victories in the 2024 Summer Olympics in Paris, France. Key moments include:
#    - [00:00:01] The Olympic rings are shown with laser lights and fireworks in the opening ceremonies.
#    - [00:00:02–00:00:08] Various shots of the games’ venues are shown, including aerial views of skateboarding and volleyball venues, a view of the track and field stadium, and a shot of the Palace of Versailles.
#    - [00:00:09–00:01:16] A fast-paced montage shows highlights from various Olympic competitions.
#    - [00:01:17–00:01:29] The video switches to show athletes celebrating victories, both tears of joy and tears of sadness are shown.
#    - [00:01:30–00:02:26] The montage then continues to showcase sporting events, including cycling, kayaking, swimming, track and field, gymnastics, surfing, basketball, and ping-pong.
#    - [00:02:27–00:04:03] More athletes celebrate their wins.
#    - [00:04:04–00:04:55] More Olympic sports are shown, followed by more celebrations.
#    - [00:04:56] Olympic medals are shown.
#    - [00:04:57] An aerial shot of the Eiffel Tower lit up with the Olympic rings is shown at night.
#    - [00:04:58–00:05:05] The video ends with a black screen and the words, “Sport. And More Than Sport.” written beneath the Olympic rings.

A seguir

Para pesquisar e filtrar exemplos de código de outros Google Cloud produtos, consulte a pesquisa de exemplos de código doGoogle Cloud .