ServiceNow, Hugging Face, and NVIDIA recently announced the launch of StarCoder2, a new set of open-access large language models (LLMs) focused on code generation. This collaboration between the companies aims to offer models that are not only high-performing but also transparent and cost-effective. StarCoder2 is a product of the BigCode community's efforts, with ServiceNow, a company specializing in digital workflows, and Hugging Face, a platform known for its extensive open-source contributions to the machine learning community, playing pivotal roles in its development.
Designed to be versatile, StarCoder2 has been trained on 619 programming languages, enabling its application across a variety of coding tasks such as generating application source codes, automating workflows, and summarizing texts. The model is structured to support further training and integration into enterprise systems for specialized functions. It is available in three sizes to suit different needs and computational capacities: a 3 billion-parameter model developed by ServiceNow, a 7 billion-parameter model from Hugging Face, and a 15 billion-parameter model created by NVIDIA using NVIDIA NeMo technology on their accelerated computing infrastructure. These options aim to balance computational efficiency with performance, offering a choice between smaller, less resource-intensive models and larger, more powerful versions.
“StarCoder2 stands as a testament to the combined power of open scientific collaboration and responsible AI practices with an ethical data supply chain,” emphasized Harm de Vries, lead of ServiceNow's StarCoder2 development team, and co-lead of BigCode. "The state-of-the-art open-access model improves on prior generative AI performance to increase developer productivity and provides developers equal access to the benefits of code generation AI, which in turn enables organizations of any size to more easily meet their full business potential.”
“The joint efforts led by Hugging Face, ServiceNow and NVIDIA enable the release of powerful base models that empower the community to build a wide range of applications more efficiently with full data and training transparency,” said Leandro von Werra, machine learning engineer at Hugging Face and co‑lead of BigCode. “StarCoder2 is a testament to the potential of open‑source and open science as we work toward democratizing responsible AI.”“StarCoder2 stands as a testament to the combined power of open scientific collaboration and responsible AI practices with an ethical data supply chain,” emphasized Harm de Vries, lead of ServiceNow's StarCoder2 development team, and co-lead of BigCode.
false
