Public record
Software health reportschema 0.34.0 · metrics 2.10.0 · 2026-09-05 22:25 UTC

pytorch / rl

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

PythonMIT★ 3,548 stars⑂ 480 forkssince Feb 2022View on GitHub ↗
KindCommand-line toolPluginLibraryhow this is determined
#ai#control#decision-making#distributed-computing#machine-learning#marl#model-based-reinforcement-learning#multi-agent-reinforcement-learning#pytorch#reinforcement-learning#rl#robotics#torch

pytorch/rl holds a health index of 91 out of 100, placing it in the Excellent band. It scores highest on Vitality (97/100) and lowest on Security (21/100). It was last updated today. A single contributor accounts for most of its recent work.

91
overall / 100
Excellent

Software health index

Metrics are grouped into weighted categories on one standardized 1–100 scale. Overall starts as their weighted mean, calibrated against the distribution of the public record so bands carry percentile meaning; when public evidence triggers the High-Risk Jurisdiction Policy, the rating is adjusted and receives an At Risk ceiling of 34.

91
Exceptional93-100The record's top tier (≈ top 5%); essentially all checked criteria met
Excellent80-92Strong across the board; minor gaps
Good65-79Healthy; gaps are limited and manageable
Moderate50-64Acceptable with notable gaps; review recommended
Weak35-49Material weaknesses across several areas
At Risk20-34Significant weaknesses; adoption warrants caution
Critical1-19Severe problems (abandoned, single-maintainer, no hygiene)
VitalityCommunity &AdoptionSustainability &GovernanceEngineeringQualitySecurityAI Readiness

Score profile

Each axis is a category. The shape matters more than the average — a healthy subject fills the whole shape, while a spike-and-crater profile means strength in one dimension is masking risk in another.

The weighted overall 77 is calibrated to 91 on the published index scale (record calibration 2026-08-02).

Ownership

pytorchOrganization · verified domain
13,606 followers70 public repossince Aug 2016

This repository is backed by an organization — shared, accountable stewardship that can outlive any single maintainer.

Package ecosystems

RegistryPackageVersionDownloads / moVersionsLast publishTags
PyPItorchrl0.13.384,7813453 days agoreinforcement-learningpytorchrlmachine-learning

Metrics by category

Vitality

Is the project alive — is code being written and are releases shipping?

97Exceptional · 21% of overall

Development activity

95Exceptional
How it's scored
36/36Push recency — last push 0 days ago
31.2/36Commit cadence — 45/52 weeks with commits
18/18Commit volume — 786 commits in the last year
0/10OpenSSF Scorecard: Maintained — no data
Inputs used
commits_last_year786
human_commit_share0.96
days_since_last_push0
active_weeks_last_year45

Release discipline

100Exceptional
How it's scored
27/27Ships releases — 35 releases published
36/36Release recency — latest release 53 days ago
27/27Release cadence — a release every ~40.2 days
0/10OpenSSF Scorecard: Signed-Releases — no data
Inputs used
releases_count35
latest_release_tagv0.13.3
releases_from_tagsno
days_since_latest_release53
mean_days_between_releases40.2

Community & Adoption

Does the project have users, downloads, attention, and a welcoming setup for contributors?

88Excellent · 17% of overall

Popularity & adoption

89Excellent
How it's scored
57.6/60Stars — 3,548 stars
22.3/25Forks — 480 forks
9.1/15Watchers — 44 watchers
Inputs used
forks480
stars3,548
watchers44
growth_stateunverified
growth_factor_pct100
growth_unverified_reasonno_history

Community health

92Excellent
How it's scored
22.5/22.5README
22.5/22.5License — recognized license (MIT)
18/18CONTRIBUTING guide
13.5/13.5Code of conduct
0/7.2Issue template
6.3/6.3PR template
Inputs used
has_readmeyes
has_licenseyes
readme_badges12
has_contributingyes
has_issue_templateno
has_code_of_conductyes
readme_badge_servicescodecov.io, github.com, shields.io
has_pull_request_templateyes

Ecosystem adoption (downloads)

82Excellent
How it's scored
65.7/80Monthly downloads — 84,781 downloads/month across pypi
0/20Registry dependents — not reported by this ecosystem
Inputs used
packagestorchrl
dependents—
ecosystemspypi
total_downloads—
monthly_downloads84,781
unverified_packages_excluded—
Excluded from scoring (no data or not applicable): Registry dependents. Remaining weights renormalized.

Sustainability & Governance

Will the project survive its people — bus factor, responsiveness, who backs it, and package upkeep?

76Good · 23% of overall

Maintainer resilience (bus factor)

34At Risk
How it's scored
9/54Bus factor — 1 contributor(s) cover half of all commits
8.4/22.5Commit distribution — top contributor authored 63% of commits
13.5/13.5Contributor breadth — 99 contributors
0/10OpenSSF Scorecard: Contributors — no data
Inputs used
bus_factor1
contributors_sampled99
top_contributor_share0.627

Issue & PR responsiveness

85Excellent
How it's scored
32.6/42Issue resolution — 78% of issues closed
26.6/30PR acceptance — 2,862/3,229 decided PRs merged
13/13Newcomer PR acceptance — 9/9 first-time contributors' PRs merged in 30d
0/15OpenSSF Scorecard: Code-Review — no data
Inputs used
merged_prs2,862
open_issues191
closed_issues666
prs_merged_7d37
prs_decided_7d40
prs_merged_30d54
prs_decided_30d58
issue_closed_ratio0.777
closed_unmerged_prs367
first_time_authors_30d2
first_time_prs_merged_30d9
first_time_prs_decided_30d9

Ownership & stewardship

100Exceptional
How it's scored
30/30Ownership backing — organization-owned
20/20Verified domain
25/25Owner reach — 13,606 followers of pytorch
25/25Track record — 70 public repos, account ~10 yr old
Inputs used
followers13,606
owner_typeOrganization
is_verifiedyes
owner_loginpytorch
public_repos70
account_age_days3,675

Package maintenance

100Exceptional
How it's scored
25/25Published & resolvable — 1 package(s) on pypi
35/35Publish recency — latest publish 53 days ago
20/20Version history — 34 published versions
20/20Not deprecated — active, not deprecated or yanked
Inputs used
packagestorchrl
ecosystemspypi
any_deprecatedno
min_days_since_publish53

Engineering Quality

Are baseline engineering and documentation practices in place?

91Excellent · 19% of overall

Engineering practices

92Excellent
How it's scored
24/24CI workflows — 30 workflow(s)
24/24Tests present
16/16Linter config — setup.cfg ([flake8])
9.6/9.6Pre-commit hooks
0/6.4.editorconfig
0/20OpenSSF Scorecard: CI-Tests — no data
Inputs used
has_ciyes
has_testsyes
has_editorconfigno
has_linter_configyes
has_precommit_configyes

Documentation

90Excellent
How it's scored
30/30README
25/25Documentation directory
15/15Documentation / homepage site — https://pytorch.org/rl
10/10Repository description
10/10Topics — 13 topics
0/10Wiki
Inputs used
topicsai, control, decision-making, distributed-computing, machine-learning, marl, model-based-reinforcement-learning, multi-agent-reinforcement-learning, pytorch, reinforcement-learning, rl, robotics, torch
has_wikino
homepagehttps://pytorch.org/rl
docs_sitehttps://pytorch.org/rl
has_readmeyes
has_docs_diryes
has_descriptionyes

Security

Are visible security and supply-chain practices strong, without unresolved high-risk jurisdiction exposure?

21At Risk · 16% of overall

Security posture

1Critical
How it's scored
0/30Security policy (SECURITY.md)
0/25Dependabot config
0/25Dependency lockfiles
0/20CodeQL workflow
Inputs used
sourcefile_signals
lockfiles—
manifestsbenchmarks/requirements.txt, docs/requirements.txt, pyproject.toml, setup.cfg, setup.py
has_codeql_workflowno
has_security_policyno
has_dependabot_configno

Dependency advisories

100Exceptional
How it's scored
35/35Direct dependencies free of known advisories — no direct dependency carries a known advisory
25/25Indirect dependencies free of known advisories — no indirect dependency carries a known advisory
0/40No advisories left outstanding — no advisory carries a publication date
Inputs used
sourceosv
advisories0
affected_packages0
assessed_packages37
unassessed_packages0
affected_by_severitynone
direct_affected_packages0
Excluded from scoring (no data or not applicable): No advisories left outstanding. Remaining weights renormalized. Matched the pypi:torchrl@0.13.3 runtime dependency closure — what installing the published package pulls in — 37 packages. Reachability is not analyzed.

AI Readiness

How well is the repo equipped to be developed and maintained with AI coding agents? Carries a deliberately small weight (4%): agent tooling is a real maintenance signal, but a repository with none can still reach 100/100.

92Excellent · 4% of overall

Agent context & guidance

100Exceptional
How it's scored
45/45Agent instructions — AGENTS.md, CLAUDE.md
15/15Machine-readable docs (llms.txt) — llms.txt served by the project website (https://pytorch.org/llms.txt)
40/40Legible commit history — 86 of 96 human commits state their intent (structured subject or explanatory body)
Inputs used
has_llms_txtyes
llms_txt_urlhttps://pytorch.org/llms.txt
legible_history_share0.896
agent_instruction_filesAGENTS.md, CLAUDE.md
agent_instruction_max_bytes13,220

Verify loop (build / test / typecheck)

89Excellent
How it's scored
18/18One-command bootstrap — Makefile, docs/Makefile
22/22Automated tests
11/11Lint / format config — setup.cfg ([flake8])
11/11Static type checking — mypy.ini
0/10Reproducible environment
10/10Demonstrated agent practice — 20 of the last 100 commits agent-authored or agent-credited
8/8Automated maintenance — 4 of the last 100 commits are automated dependency updates
0/10OpenSSF Scorecard: Pinned-Dependencies — no data
Inputs used
has_nixno
has_testsyes
lockfiles—
has_dockerfileno
typed_languageno
bootstrap_filesMakefile, docs/Makefile
has_devcontainerno
has_linter_configyes
typecheck_configsmypy.ini
agent_commit_share0.2
toolchain_manifests—
dependency_bot_commit_share0.04

Code legibility for models

78Good
How it's scored
27/45Type-checkable code — Python with type-check config (mypy.ini)
50.5/55Manageable file sizes — 71/869 source files over 60KB
Inputs used
primary_languagePython
largest_source_bytes321,561
source_files_sampled869
oversized_source_files71

Machine-readable interfaces

100Exceptional
How it's scored
0/40API schema (OpenAPI/GraphQL/proto) — not applicable to this kind of software
0/20MCP server — not applicable to this kind of software
40/40Runnable examples — examples, recipes
Inputs used
example_dirsexamples, recipes
has_mcp_signalno
api_schema_files—
interfaces_expected_of—
Excluded from scoring (no data or not applicable): API schema (OpenAPI/GraphQL/proto), MCP server. Remaining weights renormalized.

Key facts

3,548GitHub stars
99contributors
786commits, last 12 months
0days since last push
35releases
1bus factor
191open issues
PyPIpackage ecosystems

Data collection warnings

  • Star history unavailable: GitHub GraphQL error: Resource not accessible by personal access token
  • OpenSSF Scorecard did not return a usable result (killed by SIGKILL — most likely the container's memory limit; 2026/09/05 22:23:49 Warning: PATs stored in env variables GITHUB_AUTH_TOKEN and GITHUB_TOKEN differ. Scorecard will use the former.); skipping Scorecard checks

More detail

Star and fork history 0 ★ / 480 ⇿
0Stars
480Forks
35Releases

When each star and fork was added, collected from GitHub and bucketed by day. Cumulative growth sits directly above the daily additions it is made of, so the two read against each other: steady organic accretion looks nothing like an abrupt, short-lived burst. Where that difference is measurable, it is reported as growth authenticity.

0100200300400500476102022-042024-062026-09
Major 0Minor 13Patch 17

Each point covers 5 days.

Direct dependencies 7
RegistryPackageVersion constraintManifest
PyPItorch>=2.1.0pyproject.toml
PyPIpyvers>=0.2.3pyproject.toml
PyPIhoptorch>=0.1.4pyproject.toml
PyPInumpy—pyproject.toml
PyPIpackaging—pyproject.toml
PyPIcloudpickle—pyproject.toml
PyPItensordict>=0.14.1,<0.15.0pyproject.toml
All dependencies 224

Full resolved dependency set from the GitHub dependency graph: 15 direct and 209 indirect (transitive) packages. The transitive closure is complete when the repository commits a lockfile.

RegistryPackageVersionRelation
PyPIcloudpickle—direct
PyPIcloudpickle3.1.2direct
PyPIhoptorch—direct
PyPIhoptorch0.1.4direct
PyPInumpy—direct
PyPInumpy2.2.6direct
PyPIpackaging—direct
PyPIpackaging26.2direct
PyPIpyvers—direct
PyPIpyvers0.2.3direct
PyPItensordict—direct
PyPItensordict0.9.0direct
PyPItorch—direct
PyPItorch2.12.0direct
PyPItorch2.14.0.dev20260713+cu126direct
PyPIabsl-py2.5.0indirect
PyPIaccelerate—indirect
PyPIaccelerate1.14.0indirect
PyPIaccelerate1.7.0indirect
PyPIale-py—indirect
PyPIale-py0.8.1indirect
PyPIannotated-doc0.0.4indirect
PyPIanyio4.14.2indirect
PyPIattrs26.1.0indirect
PyPIautorom0.4.2indirect
PyPIautorom-accept-rom-license0.6.1indirect
PyPIbitsandbytes—indirect
PyPIbitsandbytes0.46.0indirect
PyPIcertifi2026.6.17indirect
PyPIcharset-normalizer3.4.9indirect
PyPIclick8.4.2indirect
PyPIcmake—indirect
PyPIcmake4.4.0indirect
PyPIcontourpy1.3.2indirect
PyPIcoverage—indirect
PyPIcuda-bindings12.9.7indirect
PyPIcuda-pathfinder1.5.6indirect
PyPIcuda-toolkit12.6.3indirect
PyPIcycler0.12.1indirect
PyPIdatasets—indirect
PyPIdatasets3.6.0indirect
PyPIdm-control—indirect
PyPIdm-control1.0.41indirect
PyPIdm-env1.6indirect
PyPIdm-tree0.1.10indirect
PyPIdocutils—indirect
PyPIetils1.13.0indirect
PyPIexceptiongroup1.3.1indirect
PyPIexpecttest—indirect
PyPIfilelock3.29.7indirect
PyPIfonttools4.63.0indirect
PyPIfsspec2026.6.0indirect
PyPIfuture—indirect
PyPIglfw2.10.0indirect
PyPIgym0.26.2indirect
PyPIgym-notices0.1.0indirect
PyPIgymnasium—indirect
PyPIgymnasium-robotics—indirect
PyPIh110.16.0indirect
PyPIhf-xet1.5.1indirect
PyPIhttpcore1.0.9indirect
PyPIhttpx0.28.1indirect
PyPIhuggingface-hub1.23.0indirect
PyPIhydra-core—indirect
PyPIhydra-core1.3.4indirect
PyPIhydra-submitit-launcher—indirect
PyPIhypothesis—indirect
PyPIidna3.18indirect
PyPIimageio—indirect
PyPIimmutabledict—indirect
PyPIimmutabledict4.2.1indirect
PyPIimportlib-metadata9.0.0indirect
PyPIimportlib-resources7.1.0indirect
PyPIiniconfig2.3.0indirect
PyPIipython—indirect
PyPIjax—indirect
PyPIjinja23.1.6indirect
PyPIjsonschema4.26.0indirect
PyPIjsonschema-specifications2025.9.1indirect
PyPIkiwisolver1.5.0indirect
PyPIlabmaze1.0.6indirect
PyPIlangdetect—indirect
PyPIlangdetect1.0.9indirect
PyPIlerobot—indirect
PyPIlxml6.1.1indirect
PyPImarkdown-it-py4.2.0indirect
PyPImarkupsafe3.0.2indirect
PyPImatplotlib—indirect
PyPImatplotlib3.10.9indirect
PyPImdurl0.1.2indirect
PyPImemory-profiler—indirect
PyPIminari—indirect
PyPImpmath1.3.0indirect
PyPImsgpack1.2.1indirect
PyPImujoco—indirect
PyPImujoco3.8.1indirect
PyPImujoco-torch—indirect
PyPImyst-parser—indirect
PyPInetworkx3.4.2indirect
PyPIninja—indirect
PyPIninja1.13.0indirect
PyPInltk—indirect
PyPInltk3.10.3indirect
PyPInvidia-cublas-cu1212.6.4.1indirect
PyPInvidia-cuda-cupti-cu1212.6.80indirect
PyPInvidia-cuda-nvrtc-cu1212.6.85indirect
PyPInvidia-cuda-runtime-cu1212.6.77indirect
PyPInvidia-cudnn-cu129.10.2.21indirect
PyPInvidia-cufft-cu1211.3.0.4indirect
PyPInvidia-cufile-cu121.11.1.6indirect
PyPInvidia-curand-cu1210.3.7.77indirect
PyPInvidia-cusolver-cu1211.7.1.2indirect
PyPInvidia-cusparse-cu1212.5.4.2indirect
PyPInvidia-cusparselt-cu120.7.1indirect
PyPInvidia-nccl-cu122.29.3indirect
PyPInvidia-nvjitlink-cu1212.9.86indirect
PyPInvidia-nvshmem-cu123.4.5indirect
PyPInvidia-nvtx-cu1212.6.77indirect
PyPIomegaconf—indirect
PyPIonnx—indirect
PyPIonnxruntime—indirect
PyPIonnxscript—indirect
PyPIopen-spiel—indirect
PyPIopencv-python-headless—indirect
PyPIorjson3.11.9indirect
PyPIpandas—indirect
PyPIpandas2.3.3indirect
PyPIpeft—indirect
PyPIpeft0.15.2indirect
PyPIpettingzoo—indirect
PyPIpillow—indirect
PyPIpillow12.3.0indirect
PyPIplaywright—indirect
PyPIpluggy1.6.0indirect
PyPIprocgen—indirect
PyPIprotobuf—indirect
PyPIprotobuf7.35.1indirect
PyPIpsutil—indirect
PyPIpsutil7.2.2indirect
PyPIpy-cpuinfo9.0.0indirect
PyPIpybind11—indirect
PyPIpybind113.0.4indirect
PyPIpybind11-global3.0.4indirect
PyPIpygame—indirect
PyPIpyglet—indirect
PyPIpygments2.20.0indirect
PyPIpyopengl3.1.10indirect
PyPIpyparsing3.3.2indirect
PyPIpytest—indirect
PyPIpytest9.1.1indirect
PyPIpytest-asyncio—indirect
PyPIpytest-benchmark—indirect
PyPIpytest-benchmark5.2.3indirect
PyPIpytest-cov—indirect
PyPIpytest-error-for-skips—indirect
PyPIpytest-instafail—indirect
PyPIpytest-mock—indirect
PyPIpytest-rerunfailures—indirect
PyPIpytest-timeout—indirect
PyPIpytest-timeout2.4.0indirect
PyPIpytest-xdist—indirect
PyPIpython-dateutil2.9.0.post0indirect
PyPIpytorch-sphinx-theme2—indirect
PyPIpytz2026.2indirect
PyPIpyyaml—indirect
PyPIpyyaml6.0.3indirect
PyPIray—indirect
PyPIray2.56.0indirect
PyPIreferencing0.37.0indirect
PyPIregex2026.7.10indirect
PyPIrequests—indirect
PyPIrequests2.34.2indirect
PyPIrich15.0.0indirect
PyPIrpds-py0.30.0indirect
PyPIsafetensors—indirect
PyPIsafetensors0.8.0indirect
PyPIsafety-gymnasium—indirect
PyPIscipy—indirect
PyPIscipy1.15.3indirect
PyPIsentencepiece—indirect
PyPIsetuptools—indirect
PyPIsetuptools83.0.0indirect
PyPIsetuptools-scm10.2.0indirect
PyPIshellingham1.5.4indirect
PyPIsix1.17.0indirect
PyPIsphinx—indirect
PyPIsphinx-autodoc-typehints—indirect
PyPIsphinx-copybutton—indirect
PyPIsphinx-design—indirect
PyPIsphinx-gallery—indirect
PyPIsphinx-serve1.0.1indirect
PyPIsphinxcontrib-htmlhelp—indirect
PyPIsphinxcontrib-video—indirect
PyPIsympy1.14.0indirect
PyPItenacity—indirect
PyPItenacity9.1.4indirect
PyPItiktoken—indirect
PyPItokenizers0.23.0rc0indirect
PyPItomli2.4.1indirect
PyPItorchcodec—indirect
PyPItorchvision—indirect
PyPItorchvision0.29.0.dev20260714+cu126indirect
PyPItqdm—indirect
PyPItqdm4.67.1indirect
PyPItqdm4.68.4indirect
PyPItransformers—indirect
PyPItransformers5.10.1indirect
PyPItransformers5.13.1indirect
PyPItriton3.8.0+git43422b04indirect
PyPItyper0.26.8indirect
PyPItyping-extensions—indirect
PyPItyping-extensions4.16.0indirect
PyPItzdata2026.3indirect
PyPIurllib32.7.0indirect
PyPIvcs-versioning2.2.2indirect
PyPIvllm—indirect
PyPIvllm0.26.0indirect
PyPIvmas—indirect
PyPIwandb—indirect
PyPIwandb0.19.11indirect
PyPIwrapt2.2.2indirect
PyPIxformers—indirect
PyPIxformers0.0.30indirect
PyPIzipp4.1.0indirect
Dependency advisories 0

Installing pypi:torchrl@0.13.3 pulls in 37 packages, direct and transitive: 0 carry known advisories, of which 0 are direct dependencies.

No known advisories affect the assessed dependencies.

An advisory means the version recorded in the dependency graph falls inside an advisory’s affected range. Reachability is not analysed, and the graph includes development and test pins — a finding may concern tooling rather than shipped software.

Raw JSON report machine-readable

Feedback

Spotted something off in this report, or have thoughts to share? Wrong measurements, missed tooling, ideas, questions — anything is welcome. Every message is read and gets a response.

The message is kept through sign-in.

Related records

Inspected repositories sharing catalogue tags or ecosystems with pytorch/rl. Read them side by side →

100wandb/wandbThe AI developer platform. Use Weights & Biases to train and fine-tune models, and manage models from experimentation to production.Exceptional100onnx/onnxOpen standard for machine learning interoperabilityExceptional100ray-project/rayRay is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.Exceptional98replicate/cogContainers for machine learningExceptional98ultralytics/ultralyticsUltralytics YOLO26, YOLO11, YOLOv8 — object detection, instance segmentation, semantic segmentation, image classification, pose estimation, object trackingExceptional98aws/sagemaker-python-sdkA library for training and deploying machine learning models on Amazon SageMakerExceptional

Scores are signals, not warranties. They reflect publicly visible practices on GitHub — not a code audit, and not a security guarantee.

Missing data is excluded and weights renormalized, never scored as zero. Methodology is versioned and open: metrics v2.10.0, schema v0.34.0 — full methodology · metrics wiki.

How one result sits in the wider record: aggregate statistics — PyPI.