Researchers from Petuum, MBZUAI, USC, CMU, UIUC, and UCSD have introduced LLM360, an initiative to fully open-source Large Language Models (LLMs). The project aims to make the end-to-end LLM training process transparent and reproducible, providing all training code, data, model checkpoints, and intermediate results to the community. The team has released two 7B parameter LLMs, AMBER and CRYSTALCODER, along with their training code, data, and analyses.
Read more at MarkTechPost…
