-
Notifications
You must be signed in to change notification settings - Fork 336
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Document Q&A via document loaders #35
Comments
Check out the bot builder. |
I think until there is an example of this in the examples folder, it will be hard to figure out exactly how to use it
|
As discussed in discord channel before, there'll be two ways to achieve it. The one is providing OpenAI-compatible APIs so that one could use Langchain via localhost to add documents as LLM's reference. The other is writing a library similar to Langchain in C#. Considering the time consumption of building a new library, I think the first one is a better approach. Besides, @Oceania2018 also provided a good solution since BotSharp has supported vector similarity computation now. |
I just need a simple function, that accepts a string, how to extract text from a PDF file, I'll manage, there are enough libraries, I just need to be able to insert an amount of text like an entire PDF file, and get an answer BotSharp |
any news on this issue? It would be great to be able to load documents and "interrogate" them! |
Some docs have been updated. |
An update: #226 introduces the integration for Microsoft kernel-memory, which enables adding docs as information, like pdf, txt, etc.. |
Thank you for the update. I don't understand if the files must be ingested every time the app is run or if the token from the ingestion are saved in a database for future uses. Thanks in advance! |
I think database is supported so that you won't need to create vectors for the document every time. This is actually a question about kernel-memory. :) Welcome to try this feature (the PR will be merged this weekend). |
Hi, I am just trying the example 16. It seems to me that the model is loaded twice, as you can see from the following log: ====================================================================================================== /\ \ /\ \ /\
|
Then, as you can see in the following snippet, the answer to the question continues on its own... Question: Who is the producer of Super Mario Bros movie? Answer: The producers of Super Mario Bros. Movie are Universal Pictures, Illumination, and Nintendo.
Note: These records may change as new movies are released.
I thought it could be the model, but trying the same with gpt4all I get the following: Can you please help? Thanks in advance! |
Sorry looks like this issue has been unattended for a while!
That's usually an issue with antiprompts - you need to set some "antiprompts" in the inference parameters to tell them system when to stop inferring text. For example |
Consider a typical situation where one would like to "inject" some sort of information coming from a pdf, json, xml, ect. and the user would ask questions about it.
How would we implement this using LLamaSharp? Do we need some kind of word embedding stuff that is done in LangChain?
The text was updated successfully, but these errors were encountered: