I want to upload a file that consists of text but also graphs/images. It’s a document that teaches how graphs ought to be analyzed.
And then I want the user to be able to upload an image/graph. This graph would then be analyzed based on the document’s instructions (which includes both text and images/graphs).
Am I able to do this? I know the AI would take into account the TEXT in the uploaded document, but will it also analyze and understand the graph/images in the uploaded knowledge base document? Thanks
If you put in a text file (.pdf, .txt, .csv, .doxc, etc.) it will read it as text. Images will basically be ignored.
If you put in an image file AND have configured vision-enabled chatbots correctly, then it will “look at” the image using a vision model. These vision models can read text, but that’s not their primary purpose. If you enter a file with words and images that is an image file (.png, .jpg) then it will be able to look at it. But it won’t be nearly as smart as a language model reading text.
Thanks. So if I do want to upload a document that has images, what should I do? Should I
Upload the document with text and upload the image separately and get the document to refer to the images (graphs, charts) for analysis?
Put the document into image form and upload all the text and images (graphs) as image?
Is there any other recommended way?
The text/document in my knowledge base is there to describe and show how I want the images to be analyzed, etc. So both (text and images) fit together in the knowledge base.
Then the user uploads their own image (just a simple image - graph) and I will get the AI to analyze it and give an output based on my knowledge base teaching documents (text and images)
What would you recommend is the best way I do this? Thanks.
Ah, to clarify. The vision models only work for end-user uploads. It allows end-users to upload images into the chat and the Pickaxe will understand what the image is.
There is vision/image upload in the Knowledge Base. Yet.