Thank you for flagging this @fadilikrm and apologies to you and any others on this thread experiencing this 500 error. The source of this issue was related to infrastructure on the serverless-v2 API (which powers in-app test previews of Models and Workflows, presumably what you were using). The issue has been identified and is resolved (more details below).
Can you please re-run your model in the visualize tab and/or Workflow run that you were attempting, and let me know if the error persists for you?
I attempted running a model on the Visualize tab on my end including with a new model I trained ~15min ago as well as older models and was successful, no 500 error:
if of interest, a quick note on source of issue and resolution –
The in-app API that powers model previews on the “Visualize” tab on Models and Workflow previews is being transitioned to a service called Serverless API V2 (Serverless Hosted API V2 | Roboflow Docs). Note this V2 service is distinct from Object Detection | Roboflow Docs, which was unaffected from the issue experienced on this thread and runs ~100s of millions of calls weekly. The Serverless API V2 experienced an issue with caching for inactive models (newly trained models or “cold” models), and that cache issue is resolved such that the service is functioning correctly.