Pytorch implementation of convolutional neural network visualization techniques

Pytorch implementation of various neural network interpretability methods

This repo contains Grad-CAM for 3D volumes.

Gcam is an easy to use Pytorch library that makes model predictions more interpretable for humans. It allows the generation of attention maps with multiple methods like Guided Backpropagation, Grad-Cam, Guided Grad-Cam and Grad-Cam++.

The project explores CNN architectures (ResNet vs VGG, Deformable CNNs) and 3D scene reconstruction with Tiny NeRF, including ray sampling and volume rendering. Project Year: 2025

Operational transparency for AI systems. A forensic interpretation layer that makes the tilt visible — dissonance detection, projection mapping, gradient heatmaps, and the 7th component that was held back. Designed by ChatGPT. Phantom Token by Gemini. Proyecto Estrella.