# lucidrains/alphafold2

To eventually become an unofficial Pytorch implementation / replication of Alphafold2, as details of the architecture get released

Repository: https://github.com/lucidrains/alphafold2
Canonical: https://ross.abutalabs.com/products/alphafold2
Language: Python
License: MIT
License Family: permissive
Topics: artificial-intelligence, deep-learning, protein-folding, attention-mechanism
Last push: 2022-10-29T00:34:53+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": 2101, "days_push": 1405, "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 1635, forks 264 (observed 2026-08-28T04:05:14.415400+00:00)

## What it is
An unofficial PyTorch implementation of DeepMind's AlphaFold2 protein structure prediction network, built as a translation of the official Jax code with improvements to positional encoding. It provides an attention-based model that predicts distograms and optionally backbone angles from amino acid sequences and multiple sequence alignments.

## Use cases
- predict protein contact maps and distograms from amino acid sequences in pytorch
- train an alphafold2-style attention model on protein sequence data
- replicate alphafold2 architecture for research in pytorch
- predict protein backbone angles from MSA inputs
- experiment with axial attention for protein folding

## When to choose
- you want a PyTorch implementation of AlphaFold2 instead of the official Jax/DeepMind code
- you want to modify or extend the AlphaFold2 architecture for research
- you need a pip-installable, hackable model for protein structure prediction experiments

## When to avoid
- you need the full, officially validated AlphaFold2 pipeline with pretrained weights - use DeepMind's official alphafold repository
- you need production-grade protein structure prediction with all AlphaFold2 components like Evoformer and structure module
- you need an actively developed project - the repo has not seen releases since 2022

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning
- domain: deep-learning, bioinformatics, artificial-intelligence
- platform: python
- tags: protein-folding, pytorch, alphafold2, attention-mechanism, computational-biology, gpu

## Member repositories
- lucidrains/alphafold2 (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:14.415400+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 2026-08-30T03:47:01.589193+00:00, confidence not recorded.
  - readme: https://github.com/lucidrains/alphafold2 (fetched 2026-08-28T04:05:14.415400+00:00, sha 54af256256c4)
- Data as of 2026-08-30T08:39:29.467469+00:00.