Hi all,
I’m having a pretty severe issue with the Fine-Tuning API. It’s blocking nearly all of my training job requests with the message “The job failed due to an invalid training file. This training file was blocked by our moderation system because it contains too many examples that violate OpenAI’s usage policies, or because it attempts to create model outputs that violate OpenAI’s usage policies.” Sometimes it’s the validation file instead of the training file, but its the same result regardless.
I know the system has moderation checks, but is very inconsistent in what it is blocking. All of the data I am using has previously been used in fine tuning jobs without any issues. I also had an instance where I ran two jobs at the same time with the same training and validation data, the only difference being one used the system role and the other used the developer role, and one job ran to completion and the other failed for the above reason. I think this may be a bug, but if anyone has any idea what is happening here and how to resolve it, I would be very appreciative.