GPT Plus (paid I might remember everyone) is becoming more useless by the day. It can’t even load word files with a few pages.
I encountered a similar issue with uploading files for custom GPT. In my case, the problem was caused by special Unicode characters in the text. I managed to solve it by removing all characters matching the following REGEX pattern: [\x00-\x08\x0B\x0C\x0E-\x1F\x7F]. Once these were removed, the file upload and text extraction worked flawlessly. Hope this helps!
3 Likes
Do this in Adobe and pdf file will not make any issue with ChatGpt
I was able to solve “Unable to extract text from …” by using the “Upload from computer” dialog to attach the files instead of drag and drop into a Chrome browser.
1 Like
Is there a text-to-token converter - either script or online - that we can just drop our text into - trim it at the desired length, and save it. Also, will the GPT read the file in upload order - or red a tag of some kind in the file - so that context isn’t mixed up when oder-listed is important?