chattr is a package that facilitates interaction with Large Language Models (LLMs) like GitHub Copilot Chat and OpenAI’s GPT 3.5 and 4. The primary tool is a Shiny app that operates within the RStudio IDE. See below for an example of how it appears in the Viewer pane:
“`html
“`
While this article showcases chattr’s integration with the RStudio IDE, it is important to note that it can also function outside of RStudio, such as in the terminal.
### Getting started
To begin, download the package from GitHub and launch the Shiny app using the `chattr_app()` function:
“`html
# Install from GitHub
remotes::install_github(“mlverse/chattr”)
# Run the app
chattr::chattr_app()
“`
After selecting the desired model to interact with, the app will open. A screenshot below outlines the various buttons and keyboard shortcuts available in the app:
“`html
“`
To close the app, press the ‘Escape’ key. For advanced customization, utilize the `chattr_defaults()` function. For personalized setup, chattr identifies the models you have configured and includes them in the selection menu.
For further details and instructions, visit the [chattr website](https://chattr.mlverse.org/).
—
In addition to the Shiny app, chattr offers alternative methods to engage with the LLM:
– Use the `chattr()` function
– RStudio Add-ins
For interactions with local LLMs, chattr works with LlamaGPTJ-chat. To extend chattr’s functionality, package developers can define a `ch_submit()` method in their package.
Your feedback is valuable. Feel free to share your thoughts or report issues in the [chattr’s GitHub repository](https://github.com/mlverse/chattr). Subscribe to receive notifications of new posts via email. Enjoyed this blog? Posts are also available on [r-bloggers](https://www.r-bloggers.com/).
Source link
