szagoruyko/pytorchviz
PyTorchViz renders PyTorch model computation graphs as visual diagrams for debugging and understanding model architecture.

Velocity · 7d
+1.1
★ / day
Trend
→steady
star history
This package provides make_dot and related functions that generate Graphviz visualizations of PyTorch execution traces. It shows the forward pass computational graph and optionally reveals what autograd saves for backward pass debugging. It is used by ML practitioners to inspect layer connections and data flow within neural networks.