increase number of sources?

[ram-sh]

Hello,

Awesome project! Thanks for sharing!

I had 20,000 text files embedded in my chroma db, and each file is a short story, i currently use Wizard-Vicuna-13B-Uncensored-HF as my model, and I wanted to know how to change the model to use more than just 4 sources in the creation of a new story? or perhaps I approach this the wrong way?

thanks again for your work!!
Best,
RS

[PromtEngineer]

In the retreiver object, you can set the number of documents you want to use as top_k, here is the link: https://python.langchain.com/en/latest/reference/modules/retrievers.html#langchain.retrievers.ChatGPTPluginRetriever.top_k

[LazyCat420]

Was wondering where is the retriever object located or how can you identify it? found the retriever function in run_localGPT_API.py but couldn't figure out how to get it to work.

[gulzatmamytbek1]

I would like to know this too. Good question

[PromtEngineer]

You can change the number of sources here:

localGPT/run_localGPT.py

Line 190 in 6f89470

retriever = db.as_retriever()

By replacing it with: retriever = db.as_retriever(search_kwargs={"k": 5})