How to use AzureChatOpenAI?

[IterNobody]

Seems that the component ChatOpenAI can only connect to openai?

[MatthewChapmanMIS]

This worked for me:
chat_model = AzureChatOpenAI(
openai_api_type="azure"
,model="gpt-4-32k"
,openai_api_key=""
,openai_api_base = "https://.openai.azure.com"
,openai_api_version="2023-05-15"
,deployment_name="gpt-4-32k"
)

[yassineselmi]

This worked for me as well (I use v0.4.18):

• Use the ChatOpenAI component
• Set model kwargs as:

{
    "engine": "YOUR-DEPLOYMENT-NAME",
    "api_version": "2023-05-15",
    "api_type": "azure"
}

• Set OpenAI API Base to https://YOUR-DEPLOYMENT-NAME.openai.azure.com/

[inesdp28]

That worked for me. Thank you so much. In my case the generation is being stopped - for example in a text of 10 min. requested - I got it stopped at 2 or 3 minutes of generation. You know why this could be?

[zirubak]

Hi @inesdp28 ,

I try to use AzureChatOpenAI but it doesn't work.
Do you try to use it with the latest version?

Thanks

[oguzhantortop]

Hi,
I recently downloaded LangChain but couldn't see the AzureChatOpenAI component. So I dragged the one from OpenAIChat and modified the python code like below:

from typing import Optional, Union

from langchain.llms import BaseLLM
from langchain_openai.chat_models.azure import AzureChatOpenAI

from langflow import CustomComponent
from langflow.field_typing import BaseLanguageModel, NestedDict


class ChatOpenAIComponent(CustomComponent):
    display_name = "AzureChatOpenAI"
    description = "`OpenAI` Chat large language models API."

    def build_config(self):
        return {
            "max_tokens": {
                "display_name": "Max Tokens",
                "field_type": "int",
                "advanced": False,
                "required": False,
            },
            "model_kwargs": {
                "display_name": "Model Kwargs",
                "field_type": "NestedDict",
                "advanced": True,
                "required": False,
            },
            "model_name": {
                "display_name": "Model Name",
                "field_type": "str",
                "advanced": False,
                "required": False,
                "options": [
                    "gpt-4-turbo-preview",
                    "gpt-4-0125-preview",
                    "gpt-4-1106-preview",
                    "gpt-4-vision-preview",
                    "gpt-3.5-turbo-0125",
                    "gpt-3.5-turbo-1106",
                ],
            },
            "openai_api_base": {
                "display_name": "OpenAI API Base",
                "field_type": "str",
                "advanced": False,
                "required": False,
                "info": (
                    "The base URL of the OpenAI API. Defaults to https://api.openai.com/v1.\n\n"
                    "You can change this to use other APIs like JinaChat, LocalAI and Prem."
                ),
            },
            "openai_api_key": {
                "display_name": "OpenAI API Key",
                "field_type": "str",
                "advanced": False,
                "required": False,
                "password": True,
            },
            "temperature": {
                "display_name": "Temperature",
                "field_type": "float",
                "advanced": False,
                "required": False,
                "value": 0.7,
            },
        }

    def build(
        self,
        max_tokens: Optional[int] = 256,
        model_kwargs: NestedDict = {},
        model_name: str = "gpt-4-1106-preview",
        openai_api_base: Optional[str] = None,
        openai_api_key: Optional[str] = None,
        temperature: float = 0.7,
    ) -> Union[BaseLanguageModel, BaseLLM]:
        if not openai_api_base:
            openai_api_base = "https://api.openai.com/v1"
        return AzureChatOpenAI(
            max_tokens=max_tokens,
            model_kwargs=model_kwargs,
            azure_deployment="deployment name", //change with your deployment name
            openai_api_version="2024-02-15-preview", //this is the latest version at the moment
            azure_endpoint=openai_api_base,
            api_key=openai_api_key,
            temperature=temperature,
        )

After saving the code as above you need to set your base API and API key from GUI. Then you may connect it to ConversationChain.

[Cristhianzl]

Hi! 👋

We are using other channels as our official means of communication with users. We apologize for the delayed response.
For faster assistance, please direct your questions to our Discord channel or open a GitHub issue, as we are much more active on these platforms.

Langflow Discord Server

Thank you for your understanding.

Best regards,
The Langflow Team