Deep-Learning

2025

Triton Common Pitfalls Oct 2

From the perspective of a newbie user

SnapViewer Devlog #3: Optimizations Oct 2

Intro: Troubleshooting Memory and Speed Performance

Snapviewer Devlog #2: UI Oct 2

Intro: Building the UI as a Hybrid Rust & Python Application

Notes on Writing PyTorch CUDA Extensions Oct 2

Intro: PyTorch is a Deep Learning Operating System.

近期GNN Attention算子优化工作速览 Oct 2

注：本文用LLM辅助写作的地方主要在：我认为LLM比我理解的更好的地方，会用LLM的表述代替。

SnapViewer: Faster PyTorch Memory Allocation Viewer Oct 1

When training models with PyTorch, out-of-memory (OOM) errors are common, necessitating GPU memory optimization. When simple methods (like reducing batch size) no longer work, analyzing the memory footprint of the model itself may be required.

Snapviewer Devlog #3: 性能优化 Jun 7

免责声明：我主要在 Windows 上使用最新的稳定版 Rust 工具链和 CPython 3.13 进行开发和测试。

SnapViewer: 更快的PyTorch显存分配可视化 Jun 6

在使用 PyTorch 训练模型时，内存不足（OOM）错误是很常见的，因此需要对 GPU 内存进行优化。当简单的方法（如减少batch size）不再有效时，就需要分析模型本身的内存占用情况。