Movatterモバイル変換


[0]ホーム

URL:


Skip to main content
OurBuilding Ambient Agents with LangGraph course is now available on LangChain Academy!
Open In ColabOpen on GitHub

Dall-E Image Generator

OpenAI Dall-E are text-to-image models developed byOpenAI using deep learning methodologies to generate digital images from natural language descriptions, called "prompts".

This notebook shows how you can generate images from a prompt synthesized using an OpenAI LLM. The images are generated usingDall-E, which uses the same OpenAI API key as the LLM.

# Needed if you would like to display images in the notebook
%pip install--upgrade--quiet opencv-python scikit-image langchain-community
import os

from langchain_openaiimport OpenAI

os.environ["OPENAI_API_KEY"]="insertapikey"
API Reference:OpenAI

Run as a chain

from langchain.chainsimport LLMChain
from langchain_community.utilities.dalle_image_generatorimport DallEAPIWrapper
from langchain_core.promptsimport PromptTemplate
from langchain_openaiimport OpenAI

llm= OpenAI(temperature=0.9)
prompt= PromptTemplate(
input_variables=["image_desc"],
template="Generate a detailed prompt to generate an image based on the following description: {image_desc}",
)
chain= LLMChain(llm=llm, prompt=prompt)
image_url= DallEAPIWrapper().run(chain.run("halloween night at a haunted museum"))
image_url
# You can click on the link above to display the image
# Or you can try the options below to display the image inline in this notebook

try:
import google.colab

IN_COLAB=True
except ImportError:
IN_COLAB=False

if IN_COLAB:
from google.colab.patchesimport cv2_imshow# for image display
from skimageimport io

image= io.imread(image_url)
cv2_imshow(image)
else:
import cv2
from skimageimport io

image= io.imread(image_url)
cv2.imshow("image", image)
cv2.waitKey(0)# wait for a keyboard input
cv2.destroyAllWindows()

Run as a tool with an agent

from langchain_community.tools.openai_dalle_image_generationimport(
OpenAIDALLEImageGenerationTool,
)
from langchain_community.utilities.dalle_image_generatorimport DallEAPIWrapper
from langchain_openaiimport ChatOpenAI
from langgraph.prebuiltimport create_react_agent

llm= ChatOpenAI(model="gpt-3.5-turbo", temperature=0)
api_wrapper= DallEAPIWrapper()
dalle_tool= OpenAIDALLEImageGenerationTool(api_wrapper=api_wrapper)

tools=[dalle_tool]

agent= create_react_agent(llm, tools, debug=True)

# User prompt
prompt="Create an image of a halloween night at a haunted museum"

messages=[
# "role": "user" Indicates message is coming from user
# "content": prompt is where the user's input is placed
{"role":"user","content": prompt}
]

# Sending the message to be processed and adjusted by ChatGPT, after which is sent through DALL-E
response= agent.invoke({"messages": messages})

print(response)

Related


[8]ページ先頭

©2009-2025 Movatter.jp