TylerYep/torchinfo

Health v2 · maintenance only

Adoption not part of the score

2947 stars · 138 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

torchinfo is a Python library that provides Keras-style model summaries for PyTorch, showing layer types, input/output shapes, parameter counts, and multiply-add operations. It is a rewritten successor to torchsummary with a cleaner API for inspecting and debugging neural network architectures.

Use cases

• view pytorch model summary like keras
• count parameters in a pytorch model
• inspect layer shapes of a neural network
• debug pytorch network architecture
• estimate model compute with mult-adds
• replace torchsummary with maintained alternative

When to choose

• you need a detailed, maintained model summary for PyTorch models
• you want parameter counts and FLOPs per layer for debugging
• you miss TensorFlow's model.summary() when working in PyTorch

When to avoid

• you use TensorFlow/Keras, which has model.summary() built in
• you need full computational graph visualization rather than a tabular summary
• you require support for very old Python or PyTorch versions no longer backported

Facets

library · maturity active

data-visualization developer-tools machine-learning deep-learning machine-learning developer-tools python pytorch model-summary torchsummary debugging neural-networks

Member repositories

For agents

markdown · JSON · MCP: product_card(name="TylerYep/torchinfo")