view article Article 流式数据集：效率提升 100 倍 +3 andito, lhoestq, burtenshaw, pcuenq, merve • Oct 27, 2025 • 7
StreamBP: Memory-Efficient Exact Backpropagation for Long Sequence Training of LLMs Paper • 2506.03077 • Published Jun 3, 2025 • 17
