szagoruyko/pytorchviz

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

3500 stars · 290 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

A small Python package (torchviz) that generates visualizations of PyTorch execution graphs and autograd traces using Graphviz. It renders computation graphs of model outputs, optionally showing saved tensors and backward-pass attributes.

Use cases

• visualize pytorch computation graph
• debug autograd backward pass dependencies
• see what tensors autograd saves for backward
• render neural network execution graph as image
• inspect model graph in jupyter notebook

When to choose

• you need quick Graphviz diagrams of PyTorch autograd graphs
• you're debugging gradient flow or memory usage in a PyTorch model
• you want a lightweight pip-installable visualization utility

When to avoid

• you need interactive or runtime training dashboards like TensorBoard
• you're not using PyTorch
• you need full model architecture diagrams rather than execution graphs

Facets

library · maturity maintenance

data-visualization machine-learning developer-tools deep-learning data-visualization developer-tools python cross-platform pytorch autograd computational-graph graphviz debugging

Member repositories

For agents

markdown · JSON · MCP: product_card(name="szagoruyko/pytorchviz")