Updated draft of the Deep Learning Formula Cheat Sheet.
Includes:
- Current PDF formula reference
- LaTeX source
- 20 formula sections
- tensor shape reference tables
- formula hierarchy labels: Core / Derived / Optional
- added high-value formulas: gradient checking, label smoothing, focal loss, gradient clipping, depthwise separable convolution, negative sampling, GloVe, and BLEU
- refreshed README preview images
- GitHub Actions PDF build workflow
Known notes:
- Formula conventions are documented where alternatives exist.
- Further formula verification and optional layout polish may continue in later versions.