The request of the DallE endpoint carries a parameter that defines the number of images to be generated. The sample policy will use this parameter to dynamically adjust the rate limit applied to the request.
Reference: https://learn.microsoft.com/en-us/azure/ai-services/openai/reference#image-generation