Spectral Normalisation: Why Bounding the Lipschitz Constant Stabilises Deep Networks
A deep network can be accurate on its training data and still be fragile. Change an input by an amount too small to see, and the output can swing wildly. Train a discriminator a little too hard, and the gradients explode or collapse. Spectral normalisation attacks both problems with one idea: contr…
Read the full story at DEV Community — Machine Learning ↗
Timeline · 1 report
- 2026-11-18 22:26 · DEV Community — Machine Learning
Spectral Normalisation: Why Bounding the Lipschitz Constant Stabilises Deep Networks