# aqlaboratory/openfold

Trainable, memory-efficient, and GPU-friendly PyTorch reproduction of AlphaFold 2

Repository: https://github.com/aqlaboratory/openfold
Canonical: https://ross.abutalabs.com/products/openfold
Language: Python
License: Apache-2.0
License Family: permissive
Topics: pytorch, protein-structure, alphafold2
Last push: 2025-12-16T09:41:53+00:00

## Health v2 (maintenance only)
Score: 48/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 57, release rhythm 8, longevity 100
- inputs: {"age_days": 1814, "days_push": 260, "days_rel": 495, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3420, forks 689 (observed 2026-08-28T04:08:03.797824+00:00)

## What it is
OpenFold is a faithful, trainable PyTorch reproduction of DeepMind's AlphaFold 2 for protein structure prediction. It is memory-efficient and GPU-friendly, supporting both inference and retraining with open training data (OpenProteinSet).

## Use cases
- predict 3D protein structures from amino acid sequences
- retrain or fine-tune an AlphaFold2-like model on custom data
- run protein folding inference on GPU with PyTorch
- study AlphaFold2's learning mechanisms and generalization
- predict protein complex structures with multimer modeling

## When to choose
- you need a trainable or fine-tunable AlphaFold2 implementation in PyTorch
- you want an open-source, permissively licensed alternative to AlphaFold2
- you need memory-efficient GPU inference or training for protein folding

## When to avoid
- you only need quick protein structure predictions without customization (use AlphaFold2 or ColabFold)
- you work outside PyTorch ecosystems (e.g., JAX-based workflows)
- you need pretrained weights under a fully unrestricted license (weights are CC BY 4.0)

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, llm-training
- domain: bioinformatics, machine-learning, deep-learning
- platform: python
- tags: protein-structure-prediction, alphafold2, pytorch, structural-biology, computational-biology, gpu, linux, docker

## Member repositories
- aqlaboratory/openfold (main) score 48

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:03.797824+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 2026-08-29T18:38:09.671934+00:00, confidence not recorded.
  - readme: https://github.com/aqlaboratory/openfold (fetched 2026-08-28T04:08:03.797824+00:00, sha c3060ebc9eab)
  - registry_pypi: https://pypi.org/pypi/openfold/json (fetched 2026-08-29T09:32:28.308544+00:00, sha 2949bbb3f2bc)
- Data as of 2026-08-30T08:39:29.467469+00:00.