🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

A high-throughput and memory-efficient inference and serving engine for LLMs

tiktoken is a fast BPE tokeniser for use with OpenAI's models.

Ongoing research training transformer models at scale

Community maintained hardware plugin for vLLM on Ascend

Efficient and easy multi-instance LLM serving