Eden AI

Eden AI 通过联合最佳的 AI 提供商，彻底改变了 AI 领域，使用户能够释放无限的可能性，并充分利用人工智能的真正潜力。凭借其一体化、全面且便捷的平台，它允许用户以闪电般的速度将 AI 功能部署到生产环境中，从而通过单个 API 轻松访问 AI 功能的全部广度。（网站：https://edenai.co/）

此示例介绍了如何使用 LangChain 与 Eden AI 模型交互

---

访问 EDENAI 的 API 需要 API 密钥，

您可以通过创建帐户 https://app.edenai.run/user/register 并前往此处 https://app.edenai.run/admin/account/settings 获取

获得密钥后，我们将希望通过运行以下命令将其设置为环境变量

export EDENAI_API_KEY="..."

如果您不想设置环境变量，则可以通过 edenai_api_key 命名参数直接传入密钥

在初始化 EdenAI LLM 类时

from langchain_community.llms import EdenAI

API 参考：EdenAI

llm = EdenAI(edenai_api_key="...", provider="openai", temperature=0.2, max_tokens=250)

调用模型

EdenAI API 整合了各种提供商，每个提供商都提供了多个模型。

要访问特定模型，您只需在实例化期间添加“model”。

例如，让我们探索 OpenAI 提供的模型，例如 GPT3.5

文本生成

from langchain.chains import LLMChain
from langchain_core.prompts import PromptTemplate

llm = EdenAI(
    feature="text",
    provider="openai",
    model="gpt-3.5-turbo-instruct",
    temperature=0.2,
    max_tokens=250,
)

prompt = """
User: Answer the following yes/no question by reasoning step by step. Can a dog drive a car?
Assistant:
"""

llm(prompt)

API 参考：LLMChain | PromptTemplate

图像生成

import base64
from io import BytesIO

from PIL import Image


def print_base64_image(base64_string):
    # Decode the base64 string into binary data
    decoded_data = base64.b64decode(base64_string)

    # Create an in-memory stream to read the binary data
    image_stream = BytesIO(decoded_data)

    # Open the image using PIL
    image = Image.open(image_stream)

    # Display the image
    image.show()

text2image = EdenAI(feature="image", provider="openai", resolution="512x512")

image_output = text2image("A cat riding a motorcycle by Picasso")

print_base64_image(image_output)

带回调的文本生成

from langchain_community.llms import EdenAI
from langchain_core.callbacks import StreamingStdOutCallbackHandler

llm = EdenAI(
    callbacks=[StreamingStdOutCallbackHandler()],
    feature="text",
    provider="openai",
    temperature=0.2,
    max_tokens=250,
)
prompt = """
User: Answer the following yes/no question by reasoning step by step. Can a dog drive a car?
Assistant:
"""
print(llm.invoke(prompt))

API 参考：EdenAI | StreamingStdOutCallbackHandler

链接调用

from langchain.chains import LLMChain, SimpleSequentialChain
from langchain_core.prompts import PromptTemplate

API 参考：LLMChain | SimpleSequentialChain | PromptTemplate

llm = EdenAI(feature="text", provider="openai", temperature=0.2, max_tokens=250)
text2image = EdenAI(feature="image", provider="openai", resolution="512x512")

prompt = PromptTemplate(
    input_variables=["product"],
    template="What is a good name for a company that makes {product}?",
)

chain = LLMChain(llm=llm, prompt=prompt)

second_prompt = PromptTemplate(
    input_variables=["company_name"],
    template="Write a description of a logo for this company: {company_name}, the logo should not contain text at all ",
)
chain_two = LLMChain(llm=llm, prompt=second_prompt)

third_prompt = PromptTemplate(
    input_variables=["company_logo_description"],
    template="{company_logo_description}",
)
chain_three = LLMChain(llm=text2image, prompt=third_prompt)

# Run the chain specifying only the input variable for the first chain.
overall_chain = SimpleSequentialChain(
    chains=[chain, chain_two, chain_three], verbose=True
)
output = overall_chain.run("hats")

# print the image
print_base64_image(output)

相关

• LLM 概念指南
• LLM 操作指南