OpenArch – PyTorch implementations of modern LLM architectures
OpenArch is an open‑source library that offers PyTorch code for a range of contemporary large language model designs. The project includes reference implementations of transformer‑based architectures such as GPT‑Neo, LLaMA, and others, aiming to simplify research and experimentation. Documentation covers model configuration, training pipelines, and performance benchmarks, allowing developers to reproduce and extend state‑of‑the‑art LLMs.