Hey everyone,
I’m currently in the process of fine-tuning on my dashboard, and I’m running into a bit of a problem. I’m using the training file file-Y9KyHULsGf5FwXDAPUo5Cm, which initially had a total of 9443 examples (1730 + 7713). However, 1730 examples with images were skipped for reasons like “contains children, contains faces, contains people, inaccessible URL, invalid image format”.
The core problem here is that an excessive number of images are being skipped. I’m really confused because I’m almost certain that the vast majority of these images are just plant photos. They have appropriate sizes and formats, and there’s no way they could violate OpenAI’s policies.
I’ve checked and double - checked these images, and they are simply pictures of various plants in different settings. I’m not sure why they are being flagged as containing elements like children, faces, or people. Also, the URLs are accessible, and the image formats are valid.
I’m following the process as described in the docs, but I still can’t figure out what’s going wrong. I’ve visited the docs to learn how to resolve these issues, but I haven’t found a solution that fits my situation.
Has anyone else encountered this problem? If so, could you share your experiences and how you managed to fix it? Any help would be greatly appreciated as I really want to make the most of all the examples in my training file.