Hacker News new | past | comments | ask | show | jobs | submit login
Zero-Offload: Democratizing Billion-Scale Model Training (arxiv.org)
1 point by coolspot on Feb 3, 2021 | hide | past | favorite | 1 comment



> ZeRO-Offload changes the large model training landscape by making large model training accessible to nearly everyone. It can train models with over 13 billion parameters on a single GPU, a 10x increase in size compared to popular framework such as PyTorch, and it does so without requiring any model change from the data scientists or sacrificing computational efficiency.

Official open-source implementation: https://www.deepspeed.ai/




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: