Hacker News Reader

Comment by gecersiznick | Hacker News Reader

Does anyone have any pointers as to how I can train ChatGPT on my personal corpus and create my own personal assistant?

Check OpeanAI's tutorials [1] and cookbook [2]. Or try LangChain [3], which abstracts a lot of the work.

[1] - https://platform.openai.com/docs/tutorials [2] - https://github.com/openai/openai-cookbook [3] - https://github.com/hwchase17/langchain

This is great. Thanks!

It's insanely easy to use RelevanceAI to upload a 1-column spreadsheet, click the "vectorize" button, and then have something working. The problem is putting the data into that spreadsheet.

https://relevanceai.com/blog/use-all-of-your-blog-posts-and-...

As far as the next step...

Google allows you to export your email data. Ideally, you could download that data and upload it to a service that looked at original messages and your replies (and replies to replies) and trained on that file without having to make sense of what Google gives you.

If that was working, then you could have it reply to emails for you.

Just shove all your information in the prompt. I assume that's what these guys are doing anyway. ChatGPT doesn't yet have any API, but you can use text-davinci-003, which does have an API and is functionally identical.

The API costs for a large prompt are quite high, and since you have to repeat it with every single query, your costs can go up very quickly.

Gotcha! Thanks for the response. So if I have, say, 10 PDF files which I plan to extract text from and use to fine-tune davinci, I would have to supply the entire textual information extracted from the PDF file each time?

davinci-003 has a 4k tokens cap, you cannot just 'shove all your information in the prompt.' Can parallelize but then rate limits become a problem. Really curious to hear how these guys are handling it.

You shove the whole corpus in a vector db using embeddings, query the nearest neighbors to an input and inject those into a prompt to pass to GPT.

https://langchain.readthedocs.io/en/latest/modules/chains/co...

I thought fine-tuning allows you do "inject the prompt" only once. https://platform.openai.com/docs/guides/fine-tuning

Ah, the rub is that a fine-tuned Davinci costs 6x a standard Davinci. Interesting. https://openai.com/api/pricing/

Reply on news.ycombinator.com