Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/XingjianTao/Cursor-Rules-for-PyTorch-DeepLearning-BeginnerWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/xingjiantao/cursor-rules-for-pytorch-deeplearning-beginner/pytorch-cn)<a href="https://agentmods.dev/rules/xingjiantao/cursor-rules-for-pytorch-deeplearning-beginner/pytorch-cn"><img src="https://agentmods.dev/badge/rules/xingjiantao/cursor-rules-for-pytorch-deeplearning-beginner/pytorch-cn.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.05285 | $0.05285 |
| Opus 5 | $0.02642 | $0.02642 |
| Sonnet 5 | $0.01057 | $0.01057 |
| Haiku 4.5 | $0.00528 | $0.00528 |
Grade A, and why
pytorch-cn scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 384 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PyTorch 最佳实践和编码标准
本文档为PyTorch项目开发提供了全面的指导方针,涵盖了代码组织、性能优化、安全性、测试方法和常见陷阱。遵循这些最佳实践将使PyTorch代码更加可读、可维护和高效。
库信息:
- 名称:PyTorch
- 类别:ai_ml
- 子类别:machine_learning
1. 代码组织和结构
1.1 目录结构最佳实践
一个组织良好的目录结构可以提高代码的可维护性和协作性。以下是PyTorch项目的推荐结构:
project_root/ ├── data/ │ ├── raw/ │ ├── processed/ │ └── ... ├── models/ │ ├── layers.py │ ├── networks.py │ ├── losses.py │ ├── ops.py │ └── model_name.py ├── src/ │ ├── data/ │ │ ├── datasets.py │ │ ├── dataloaders.py │ │ └── transforms.py │ ├── models/ │ │ └── ... (模型相关代码) │ ├── utils/ │ │ └── ... (工具函数) │ └── visualization/ │ └── ... ├── notebooks/ │ └── ... (用于实验的Jupyter笔记本) ├── tests/ │ ├── unit/ │ ├── integration/ │ └── ... ├── scripts/ │ └── train.py │ └── eval.py ├── configs/ │ └── ... (配置文件,例如YAML) ├── README.md ├── requirements.txt ├── .gitignore └── ...
data/:存储原始和处理过的数据集。models/:包含PyTorch模型定义、层和自定义损失函数。将网络架构、单独的层/块和操作分为不同的文件。src/:保存主要源代码,包括数据加载、模型定义、工具函数和可视化工具。通常根据职责进一步拆分src/文件夹。notebooks/:用于实验和探索的Jupyter笔记本。使用笔记本进行初步探索和原型开发,但将最终确定的代码转移到Python脚本中。tests/:单元、集成和端到端测试。scripts/:训练、评估和部署脚本。主训练脚本应导入模型定义。configs/:用于超参数设置和其他参数的配置文件。
1.2 文件命名约定
- 使用描述性和一致的文件名。
- Python文件:
lower_with_under.py(例如,data_loader.py、model_utils.py)。 - 模型文件:
model_name.py(例如,resnet.py、transformer.py)。 - 配置文件:
config_name.yaml(例如,train_config.yaml)。
1.3 模块组织
- 将相关函数和类分组到模块中。
- 使用清晰简洁的模块名称。
- 在每个模块开头包含一个docstring来描述其用途。
- 遵循一致的导入风格:
# 标准库导入 import os import sys # 第三方库导入 import numpy as np import torch import torchvision # 本地应用程序/库导入 from src.data import data_loader from src.models import resnet from src.utils import helper_functions
1.4 组件架构
nn.Module:创建PyTorch神经网络的基本构建块。始终继承nn.Module来定义模型、层和自定义操作。forward()方法:在forward()方法中实现模块的前向传递。PyTorch使用__call__方法(它会执行forward())来通过模型传递数据。- 关注点分离:将模型设计为更小、可重用模块的组合。这促进了模块化并简化了调试。
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 384 lines · 5,285 tokens per session scan A e5a44e406093
pytorch-cn is a cursor rule published in the GitHub repository XingjianTao/Cursor-Rules-for-PyTorch-DeepLearning-Beginner (3 stars, last pushed 1y ago), licensed MIT. It adds 5,285 tokens to every session, about $0.0264 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other cursor rules, from other repositories
llm-layer
LLM provider implementation patterns.
tensorflow
TensorFlow: Keras, model training, production deployment.
006_Program_of_Thought_Tutorial
DSPY 3 Program of Thought Tutorial - Production code reasoning system from official DSPy 3.0.1 tutorial.
standards-data-eng
Mandatory standards for Python and SQL data pipelines.
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.
angular-20
This rule provides comprehensive best practices and coding standards for Angular development, focusing on modern TypeScript, standalone components, signals, and performance optimizations.