Exploring Stability AI's New Open Source Language Model (StableLM)

In this video, I explore Stability AI's new open-source language model, StableLM.
StableLM is a large language model trained on a massive dataset, optimized for various NLP tasks.
I show you how to load and run StableLM in a Colab notebook, generate text with different temperature settings, and solve programming problems like calculating the Nth Fibonacci number.
StableLM is lightweight, running on under 11GB of GPU RAM, and the models and datasets are available under Creative Commons licenses.
🔗 StableLM Blog Post: stability.ai/b...
🔗 StableLM Repo: github.com/sta...
🔗 StableLM Colab: colab.research...
#os #opensource #stability #stabilityai #llm #stablelm

[@YouJLT]

Fibonacci of course!😅

[@chrisalexiuk]

Of course! It had to be!

[@DrewIsFail]

What does the temperature have to do with it?

[@chrisalexiuk]

In the context of getting a Python function that calculates the Fibonacci sequence, choosing a lower temperatue (which gives us token sequences that are less "original"), it gives us a result more in line with the "usual" or "standard" answer to the query.
When the temperature was high, the model started improving - and was unable to do so in a way that provided an accurate result.
I hope that explains your question - and I'm sorry if I misinterpreted it!