This is the default state of deep learning projects, everyone assumes only phd researchers will ever try it who already know how to use everything in the tool c

This is the default state of deep learning projects, everyone assumes only phd researchers will ever try it who already know how to use everything in the tool chain. What's happened with llama and other LLMs with codebases that actually work outright with one click when compiled is a pretty big outlier.