Finetuning & Knowledge Retrieval

The pk.Knowledge class makes it easy to finetune a model to a specific set of documents. You can do this via a file-upload API, or a glob of local paths.

Finetune With File Upload UI

To finetune an existing model via a file-upload UI, you can run the following:

import pykoi as pk

# assume you have some model, endpoint, or api
model_1 = your_model

# give interactive chat ui with feedback
chatbot = pk.Compare(model=model)
fileupload = pk.FileUpload(model=model)

# Create sharable link to the application
app = pk.Application(debug=False, share=False)
app.add_component(chatbot)
app.add_component(fileupload)
app.run()

Which will launch:

img

This is a useful option when you want to provide users with models that they can finetune for themselves.

Finetune via Code

In addition, you can finetune your model on local documents and save it for later use. Alternatively, you can finetune a model by specifying a set of local files:

import pykoi as pk

# assume you have some model, endpoint, or api
model_1 = your_model

# give interactive chat ui with feedback
chatbot = pk.Compare(model=model)
fileupload = pk.FileUpload(model=model)

# Create sharable link to the application
app = pk.Application(debug=False, share=False)
app.add_component(chatbot)
app.add_component(fileupload)
app.run()

Finetuned models can be saved for further use elsewhere, or shared via pykoi’s chat UI’s:

Let’s see a quick example where we allow one to upload their own documents in one page, and interact with the new model in another:

Visualizing Your Finetuned Model

To understand your documents, you may use the

Local Files: