Hello everyone,
I’ve been experimenting with CustomGPT and have primarily used PDFs to upload my data. So far, I’ve condensed approximately 150 PDFs into 3 large files for this purpose. While this method has been somewhat effective, I’ve noticed two main issues that I’m hoping to optimize: upload speed and the handling of screenshots within the PDFs. The problem here lies in the limitation of 20 documents otherwise i would upload all of them single.
Given the nature of PDFs, especially those heavy with screenshots, I suspect there might be more efficient formats or strategies to prepare and upload data to CustomGPT. The goals are to enhance processing speed and ensure the screenshots and visual data are effectively utilized or managed.
- Speed: The upload and processing time for large PDFs can be quite slow. Is there a more efficient file format that CustomGPT processes faster?
- Screenshots and Visuals: A significant portion of my PDFs contains screenshots. I’m concerned about how CustomGPT utilizes this visual information, if at all. Is there a recommended practice for including visual data or should I focus on text-based data only?
- File Format Recommendations: What are the best file formats for uploading to CustomGPT, considering both text and visual content? Are there specific formats that are known to be more compatible or result in faster processing times?
- Preparation Tips: Are there any tips for preparing data before upload to optimize the CustomGPT’s learning from it? This could include how to handle multiple documents, whether to merge them, and how to deal with visual content.
I appreciate any insights, experiences, or recommendations you could share. My goal is to make the data upload process more efficient and ensure the best use of CustomGPT’s capabilities, especially when dealing with a mix of text and visuals.
Thank you!