# pytorch/torchdynamo

A Python-level JIT compiler designed to make unmodified PyTorch programs faster.

Repository: https://github.com/pytorch/torchdynamo
Canonical: https://ross.abutalabs.com/products/torchdynamo
Language: Python
License: BSD-3-Clause
License Family: permissive
Archived: true
Last push: 2024-04-17T23:15:30+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 1825, "days_push": 868, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1078, forks 128 (observed 2026-08-28T04:03:29.925038+00:00)

## What it is
TorchDynamo is a Python-level JIT compiler that speeds up unmodified PyTorch programs by capturing Python bytecode into FX graphs. The project has been merged into PyTorch itself as torch._dynamo, and this standalone repository is archived.

## Use cases
- speed up unmodified pytorch models with a jit compiler
- capture pytorch python code into fx graphs for optimization
- accelerate deep learning training and inference without code changes
- compile pytorch models ahead of deployment for faster execution

## When to choose
- you want to accelerate existing PyTorch code with torch.compile in modern PyTorch releases
- you need graph capture of Python bytecode for ML compiler research

## When to avoid
- you are starting a new project - use torch._dynamo / torch.compile bundled with PyTorch instead
- you need a standalone pip-installable torchdynamo package with active support

## Facets
- artifact type: library
- maturity: maintenance
- function: compiler, machine-learning, deep-learning, llm-training
- domain: machine-learning, deep-learning, compilers, performance
- platform: python, cross-platform
- tags: jit-compiler, pytorch, graph-capture, merged-into-pytorch, archived-repo, gpu

## Member repositories
- pytorch/torchdynamo (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:29.925038+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 2026-08-30T06:53:11.794857+00:00, confidence not recorded.
  - readme: https://github.com/pytorch/torchdynamo (fetched 2026-08-28T04:03:29.925038+00:00, sha ea69daca994b)
  - registry_pypi: https://pypi.org/pypi/torchdynamo/json (fetched 2026-08-29T12:54:34.667590+00:00, sha e3a452ad7d9e)
- Data as of 2026-08-30T08:39:29.467469+00:00.