SMI-training-dynamics
Project status: UnknownReimplementation of the Sliced Information Plane (SIP) framework from Wongso, Ghosh, and Motani (2025) for analyzing deep neural network training dynamics. The repo uses Sliced Mutual Information (SMI) to obtain scalable, finite dependence estimates in high‑dimensional, deterministic settings, and applies them to MNIST MLP experiments.
- Target software or hardware
- Work type
- Primary public repository
- github.com/fynngerding/smi-training-dynamics
Projects are sorted by the latest recorded repository star count. Stars indicate interest, not completion. Missing counts remain unknown.
Repository metadata needs refreshing. The cached listing remains available.
About and scope
Reimplementation of the Sliced Information Plane (SIP) framework from Wongso, Ghosh, and Motani (2025) for analyzing deep neural network training dynamics. The repo uses Sliced Mutual Information (SMI) to obtain scalable, finite dependence estimates in high‑dimensional, deterministic settings, and applies them to MNIST MLP experiments.
Contribution availability
Contribution policy unknown
- AI-assisted contribution policy
- Unspecified
Documented contributor systems
Contributor setup not documented
Reported progress
Progress not reported
Sources, builds and releases
- Open repository · github.comgithub.com/fynngerding/smi-training-dynamics