# DiffEqML/torchdyn

A PyTorch library entirely dedicated to neural differential equations, implicit models and related numerical methods

Repository: https://github.com/DiffEqML/torchdyn
Canonical: https://ross.abutalabs.com/products/torchdyn
Homepage: https://torchdyn.org
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Topics: deep-learning, neural-network, neural-differential-equations, pytorch, dynamical-systems, deep-equilibrium-models, implicit-models, control-theory, neural-ode, numerical-methods
Last push: 2024-05-02T02:44:43+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 2320, "days_push": 853, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1578, forks 135 (observed 2026-08-28T04:05:06.270887+00:00)

## What it is
Torchdyn is a PyTorch library dedicated to numerical deep learning, providing tools for neural differential equations, implicit models, and related numerical methods. It offers classes like NeuralODE and NeuralSDE, a functional API for GPU-compatible numerical solvers, and extensive tutorials.

## Use cases
- train neural ODEs in PyTorch
- build continuous-depth neural network models
- implement neural SDEs for stochastic dynamics
- model dynamical systems with deep learning
- solve implicit deep equilibrium models
- apply numerical methods to deep learning research
- benchmark neural differential equation solvers

## When to choose
- you need neural differential equations integrated with PyTorch
- you want GPU-compatible numerical solvers with a functional API
- you are researching continuous-depth models, neural ODEs, or deep equilibrium models
- you want tutorials and benchmarks for numerical deep learning

## When to avoid
- you need a framework-agnostic solution outside PyTorch
- you only need classical ODE solvers without deep learning integration
- you require production-hardened, long-term-stable APIs for critical systems

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, simulation, math
- domain: deep-learning, machine-learning, simulation
- platform: python, cross-platform
- tags: neural-ode, neural-differential-equations, pytorch, dynamical-systems, deep-equilibrium-models, numerical-methods, implicit-models, control-theory, algorithms, gpu

## Member repositories
- DiffEqML/torchdyn (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:06.270887+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:56:53.606156+00:00, confidence not recorded.
  - readme: https://github.com/DiffEqML/torchdyn (fetched 2026-08-28T04:05:06.270887+00:00, sha 8838d498c95b)
- Data as of 2026-08-30T08:39:29.467469+00:00.