Stability-AI/StableLM ? reverse-engineered prompt

Build me a clean public repo for a language model project like StableLM, with a clear README that explains the different model releases, what each one is for, and where to get the checkpoints. I want it to feel like an official release page, with sections for model sizes, training details, evaluation results, and links to the weights.

Please also add a few simple notebooks that show how to load a checkpoint, try a basic prompt, and run an evaluation example. Include a config file for the main model setup, plus a place to store eval outputs and any supporting assets. Keep it polished and easy to follow for someone who just wants to understand the model and use it without digging through a bunch of code.

If you need current library docs or best practices, look them up online and keep the repo organized around documentation and reproducible examples.