We at Run:ai are dedicated to make the life of data scientists and researchers easier. It’s not a secret that large language models (LLMs) are getting loads of attention and a recent survey showed that many organizations will be deploying LLMs in product within the next 12 months.

‍

Not all organizations are going to train their own LLMs from scratch but take an existing (pre-trained) model and start tailoring it to their needs. There are several ways to tailor the models but one of them, and that one is gaining some traction, is in-context learning. In-context learning basically learns the LLM to solve a new task at inference time by feeding it prompts that contain examples of those specific tasks.

‍

Now back to Run:ai, like I said before we aim to make the life of data scientists, researchers and prompt engineers easier. Our R&D is continuously working to achieve just this, one of our most recent internal alpha features was too good not to share and shows you a glimpse into what you will be able to see in our Run:ai Atlas platform. This example shows you a very easy way to deploy a model (could be any model including GenAI models) from Hugging Face and add your favorite tool to interact with it (in this case Gradio), it basically creates a “playground” for you and your team members to experiment with any type of model in a matter of seconds.

‍

This is just one of the cool things that we are working on internally to ensure we enable organizations to innovate faster.

‍

Watch the demo right here!