Skip to content
This repository has been archived by the owner on Oct 5, 2023. It is now read-only.

[DOC] re-training on custom data q #262

Open
staccDOTsol opened this issue Jul 23, 2020 · 1 comment
Open

[DOC] re-training on custom data q #262

staccDOTsol opened this issue Jul 23, 2020 · 1 comment

Comments

@staccDOTsol
Copy link

📖 Document

hey mate,

GPT-3 must be exciting for you folks at the moment :)

hey, I already scraped a story site I'd like to try recreating AI Dungeon with.

I did try doing this already, trained it up, but the results weren't really sensical.

I'm willing to bet my issue is translating the scraped data into your json dict.

Can you provide more examples here:

{
"tree_id":
"story_start":
"action_results": [
{"action":, "result":, "action_results": },
{"action":, "result":, "action_results": }]
}

Maybe by filling in the blanks?

What's the point of action_results within action_results?

@staccDOTsol
Copy link
Author

Much, much scrape. Check out scrollbar! I guess I only have even numbers for whatever reason, 4k+ stories!

image

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant