Hacker News new | ask | show | jobs
Directly run and investigate Llama models locally with only PyTorch (github.com)
3 points by anordin95 617 days ago
1 comments

There are other popular ways to invoke these models, such as Ollama and Hugging-Face's general API package: transformers, but those hide the interesting details behind an API. Peel back the layers to poke, prod and understand!