OpenAI is launching a research challenge aimed at attracting skilled engineers and researchers while advancing efficient AI development. Participants are tasked with creating a pretrained model that meets specific constraints: a 16MB size limit and a 10-minute training period using eight H100 GPUs. The competition, hosted on GitHub, encourages innovation through an open-source approach by allowing participants to improve a baseline model, submit their work, and compete for a spot on a public leaderboard. Outstanding submissions may lead to job interviews and public recognition. To assist participants, OpenAI has partnered with Runpod to offer compute credits, with options ranging from quick-start allocations to advanced grants, enabling extensive experimentation. Eligibility for these credits requires applicants to provide details like their GitHub username and role, ensuring compliance with legal restrictions. This initiative not only seeks technical advancements in AI but also serves as a talent acquisition strategy.
Source link
