Category: systems
-
CS336 Lecture 2：PyTorch 与 Resource Accounting
TLDR: Before training a model, PyTorch tensors, memory, FLOPs, and profiling have to become concrete enough that architecture choices have real resource prices.
-
AMP：Automatic Mixed Precision 技术笔记
TLDR: AMP speeds training and reduces memory by choosing lower precision for safe ops while keeping higher precision where numerical stability matters.
-
Claude Code Source：把 Agent 看成一个操作系统进程
TLDR: Looking at Claude Code as an OS process exposes the practical substrate of agents: files, permissions, plugins, subprocesses, and tool calls.