I am not able to perform inference on images after adding data sources to Azure GPT Vision models (GPT-4). Previously, I was able to do it.

Sreeja B 0

I set up RAG using files available in Azure Blob Storage and did the same with Azure Search. However, after adding data sources, I am not able to provide an image as input for inference. I am getting the following error: openai.BadRequestError: Error code: 400 - {'error': {'requestid': '****************', 'code': 400, 'message': 'Invalid chat message detected: message content must be string'}}

I am using Azure GPT-4o. It was working last month. Now, in the chat playground as well , if we attach a data source, the attachment icon is missing in the chat window.

YutongTie-MSFT 50,856 Reputation points

2024-08-28T21:45:13.2133333+00:00

Hello Sreeja,

Thanks for reaching out to us, could you please share the region you are working on? Are you working on AI Studio or Azure OpenAI Studio?

Regards,

Yutong
Sreeja B 0 Reputation points

2024-09-03T04:53:46.1833333+00:00

Region is eastus . I am using azure open ai studio . Tried with python code as well .Same code which was working earlier is now throwing below error . 'code': 400, 'message': 'Invalid chat message detected: message content must be string'}} . Now after adding data source , its not accepting image input .

Share via

I am not able to perform inference on images after adding data sources to Azure GPT Vision models (GPT-4). Previously, I was able to do it.

Your answer