Need Help with AutoGen Workflow Issue!

[Nalin Nishant (nalindev)]

Hi everyone,

I’m trying to use AutoGen to generate images but ran into a problem. Here’s what I did:

1. Set up AutoGen in a Conda virtual environment (latest version).
2. Created a GPT-4O model with a valid API key and used the default image generation skills.
3. Made a Painter Assistant using the GPT-4O model with default image generation skills.
4. Built a workflow where:
   • Initiator: user_proxy (default agent).
   • Receiver: Painter Assistant.
5. Ran the workflow.

The issue? Instead of getting a DALL-E 3 image, I’m only getting Python code as the output. 🤔

Has anyone else faced this or knows how to fix it? Would really appreciate some guidance!

Thanks! 😊

[Eric Zhu (ekzhu)]

If you post your code, we don't have to guess your issue :)

Also, looks like you are still using v0.2 API. The latest stable release is v0.4.

Migration guide: https://microsoft.github.io/autogen/stable/user-guide/agentchat-user-guide/migration-guide.html