Basically I’m trying to understand a repository from GitHub, the whole repo is about 7 MB, and I’m curious as if there is a way to download the repo and upload it to some AI (e.g. ChatGPT) that can explain the code to me and maybe change a few things.
Since the whole code has a lot of subfiles I can’t just copy and paste it, and I really don’t want to ask for help just in specific parts of the code. I’d appreciate any guidance in this.
That sounds great. Have you found a way yet?
One method that might work is turning all the code into a PDF and using the AskYourPdf plugin.
Or that might be a method that doesn’t work.
This question is like “how do I upload Github to a Commodore 64?” (even a scientific calculator may have more storage) - and then have it use all of that data all at once?
The answer is you don’t. You insert game disc 3 of 8 to play the next part of the adventure.
Or in the case of a language model that has about 2000-3000 word storage for both the input and required by the output, you get the small chunk of data that is able to answer the question, by using some other search algorithm based on the user’s question, and place the best information into the AI’s memory to enhance capabilities.
Keywords: embeddings semantic vector database augmented retrieval