Yesterday, ChatGPT would allow for PDFs and Word docs to be uploaded. Today, there is an error message for uploading. What is wrong with the service?
Same issue here. Specifically I see the following error in the console while attempting to upload an ArXiv paper:
{"file_id":"file-G4xALsnNNm5jYj1sXDj2VT","event":"file.indexing.error","message":"Error attaching files 2312.06211v2.pdf for retrieval: Value after * must be an iterable, not int.","progress":null,"extra":null}
same issue here tried word and pdf doesnt work
Interesting. For now, my error is “Unknown error occurred.”
That’s just what the toast notification says. If you check the response from the network request, it’ll have a few JSON lines with one of them detailing the error like I posted. Looks like they’re running a python sever on the backend and an error crept in.
Hopefully the poor dev that pushed code changes to prod before the holiday (or the one on call to handle issues) sees the error message I posted and it helps them find the bug quickly.
No seriously, it’s concern many people apparently with PDF and WORD files, what is going on ?
{"file_id":"file-RZEr4BCgkukLMNmcPsStmP","event":"file.processing.started","message":"Start processing file: ******.pdf","progress":0.0,"extra":null}
{"file_id":"file-RZEr4BCgkukLMNmcPsStmP","event":"file.processing.file_ready","message":"File file-RZEr4BCgkukLMNmcPsStmP ******.pdf is ready to download","progress":20.0,"extra":null}
{"file_id":"file-RZEr4BCgkukLMNmcPsStmP","event":"file.indexing.error","message":"Error attaching files ******.pdf for retrieval: Value after * must be an iterable, not int.","progress":null,"extra":null}
This is not acceptable in a production environment, especially given the cost associated with this service. The error suggests a basic issue with type handling (Value after * must be an iterable, not int.
), which should not occur in a live system.
Happening to me as well. Just started. Trying to upload a PDF to my GPT’s knowledge. I got “unknown error occurred” Damn, that blows!
It still works on the mobile app FWIW. I uploaded the needed PDF there then went back to my web client for further work with it.
Same here, getting an “Unknown error occured”.