Vertex Ai Custom Model

  • 10 December 2023
  • 3 replies

Badge +1


I am exploring Vertex AI : Chat AI package. It works perfectly fine with default models. Then I started exploring custom model. 

To explore it, I have published model in Google Cloud and test model in GCP. It is working fine. I started configuring parameters in A360 package and when I am running the BOT it is throwing me bad request 400 Invalid Argument error.


Please guide me if I am missing anything.





Best answer by Micah.Smith 12 December 2023, 17:20

View original

3 replies

Userlevel 5
Badge +9

I havent tried this specifically myself - but I would say as a first start, see if you’re able to connect to your GCP model in Postman or the REST package. 400 error may indicate the path isnt correct or something in the request isnt formed correctly - so using Postman can help you identify if its an issue with your request vs the GCP model you’re hosting and your ability to connect to it

Badge +1

Thank you for your suggestion. I am able to connect to custom model using postman. Please guide me on this as this is urgent for me to proceed with one of the use case.

Userlevel 5
Badge +9

I would then opt to use the REST package in Automation Anywhere. It sounds like the Generative AI package is only set up to work with default models, not the REST package would be your best bet - replicating what you’ve got working in Postman