SAP systems always put the charset in the content-type header. It would be great if the NVIDIA LLM APIs accepted the content-type header with the charset information. All other LLM APIs I’ve been working with don’t return error because of the charset in the content type.
Thank you!