Good zero-to-hero resources/repos on LLM training and inference?
I'm looking on resources to learn more about training, refinement and deployment of large language models.
Karpathys stuff (https://karpathy.ai/zero-to-hero.html) is a pretty good overall intro. But for the hero part, you aren't going to get that from any course material, you are going to get it from implementing and training said models, which is somewhat unfeasible due to the hardware required to train these at scale.