📄 中文摘要
PyTorch 2.11 版本的发布带来了多项重要更新，包括用于分布式训练的可微分集合（Differentiable Collectives）和 FlashAttention-4 的增强功能。新版本旨在提升模型训练的效率和性能，支持更复杂的计算需求。用户可以利用这些新特性来优化其深度学习工作流程，进一步推动研究和应用的发展。PyTorch 2.11 还包含了一些性能改进和 bug 修复，确保用户在使用时获得更稳定的体验。
PyTorch 2.11 发布博客
发布: 2026年3月23日
📄 English Summary
PyTorch 2.11 Release Blog
The release of PyTorch 2.11 introduces several significant updates, including Differentiable Collectives for distributed training and enhancements to FlashAttention-4. This new version aims to improve the efficiency and performance of model training, supporting more complex computational needs. Users can leverage these new features to optimize their deep learning workflows, further advancing research and applications. Additionally, PyTorch 2.11 includes performance improvements and bug fixes, ensuring a more stable experience for users.