How to use GPT 4 api to parse PDFs that contains graphical data


I’m trying to provide GPT-4 API with some PDF files, some of these files contain graphs. I’m currently using Ada-002 to create embeddings for text data and store them in Pinecone. Any recommendations on how to achieve this?


There is the CLIP, an OpenAI Open source (GitHub - last release Feb 20, 2023) described as: “A Contrastive Language-Image Pre-Training neural network trained on a variety of image/text pairs” - to be installed locally. It is an idea to take a look at it.