Neural Network Neural Network Victorzhou Intro Neural Networks

---
name: neural-network-neural-network-victorzhou-intro-neural-networks
description: |
  Trigger when the user wants to learn, implement, debug, or extend "从零实现神经网络" or a 神经网络 build-your-own project using Python. Use this skill for staged Claude Code guidance, milestone planning, minimal implementation, and troubleshooting.
---

# 从零实现神经网络

Use this skill as a project coach for the original build-your-own tutorial.

## Source

- **Category**: 神经网络
- **Language / stack**: Python
- **Original resource**: https://victorzhou.com/blog/intro-to-neural-networks/

## Workflow

1. Establish whether the user starts from an empty directory, a partial implementation, or a failing project.
2. Default to `Python` unless the user chooses another stack.
3. Read `references/tutorial-map.md` and convert the source into 4-8 implementation milestones.
4. Implement one milestone at a time with small patches and explicit verification commands.
5. Prefer hints, tests, and minimal fixes over dumping a complete final solution.
6. Preserve the user's existing files and style.

## Milestone template

1. Project skeleton and first runnable command.
2. Core data model and invariants.
3. Input/parser/protocol/UI boundary.
4. Main algorithm or runtime engine.
5. Integration and end-to-end behavior.
6. Diagnostics, tests, and extensions.

## Expected outputs

- A staged implementation plan.
- A minimal project scaffold.
- A focused code patch for the current milestone.
- A debugging explanation and regression test.