A SOTA quantization algorithm for high-accuracy low-bit LLM inference, seamlessly optimized for CPU/XPU/CUDA, with multi-datatype support and full compatibility with vLLM, SGLang, and Transformers.
intel/auto-round 的健康指数为 100 分中的 97 分,处于「卓越」区间。 其得分最高的类别是Vitality(99/100),最低的是AI Readiness(60/100)。 最近一次更新在今天。 近期的大部分工作由 3 位贡献者完成。
指标归入加权类别,统一采用 1–100 量表。总体分先取类别加权平均,再依据公开记录的分布进行校准,使各等级具有百分位含义;当公开证据触发高风险司法辖区政策时,评级会按政策调整,并设置 34(存在风险)的上限。
每条轴代表一个类别。形状比平均值更重要——健康的对象会填满整个图形,而“一峰一谷”式画像意味着某一维度的优势正掩盖另一维度的风险。
加权总体分 85 经校准后在公布的指数量表上为 97(记录校准 2026-08-02)。
项目是否仍有生命——是否仍在编写代码,是否仍在发布版本?
| 36/36 | 推送新近度 — 最近一次推送于 0 天前 |
| 35.3/36 | 提交节奏 — 52 周中有 51 周有提交 |
| 18/18 | 提交量 — 最近一年 758 次提交 |
| 10/10 | OpenSSF Scorecard:Maintained — 30 commit(s) and 0 issue activity found in the last 90 days -- score normalized to 10 |
| commits_last_year | 758 |
| human_commit_share | — |
| days_since_last_push | 0 |
| active_weeks_last_year | 51 |
| 27/27 | 有发布版本 — 已发布 37 个发布版本 |
| 36/36 | 发布时效 — 最近一次发布版本于 2 天前 |
| 27/27 | 发布节奏 — 约每 16.6 天发布一次 |
| 0/10 | OpenSSF Scorecard:Signed-Releases — 无数据 |
| releases_count | 37 |
| latest_release_tag | v0.14.2 |
| releases_from_tags | 否 |
| days_since_latest_release | 2 |
| mean_days_between_releases | 16.6 |
项目是否拥有用户、下载量与关注度,并具备欢迎贡献者参与的配置?
| 51.6/60 | 星标 — 1,520 个星标 |
| 18.2/25 | 复刻 — 155 个复刻 |
| 6.7/15 | 关注者 — 17 位关注者 |
| forks | 155 |
| stars | 1,520 |
| watchers | 17 |
| growth_state | unverified |
| growth_factor_pct | 100 |
| growth_unverified_reason | no_history |
| 22.5/22.5 | README |
| 22.5/22.5 | 许可证 — 可识别的许可证(Apache-2.0) |
| 18/18 | CONTRIBUTING 指南 |
| 13.5/13.5 | 行为准则 |
| 0/7.2 | 议题模板 |
| 6.3/6.3 | PR 模板 |
| has_readme | 是 |
| has_license | 是 |
| readme_badges | — |
| has_contributing | 是 |
| has_issue_template | 否 |
| has_code_of_conduct | 是 |
| readme_badge_services | — |
| has_pull_request_template | 是 |
项目能否在其成员之外延续——巴士系数、响应能力、由谁支持,以及软件包的维护状况?
| 36/54 | 巴士系数 — 3 位贡献者贡献了半数提交 |
| 15.5/22.5 | 提交分布 — 头号贡献者编写了 31% 的提交 |
| 13.5/13.5 | 贡献者广度 — 37 位贡献者 |
| 6/10 | OpenSSF Scorecard:Contributors — project has 2 contributing companies or organizations -- score normalized to 6 |
| bus_factor | 3 |
| contributors_sampled | 37 |
| top_contributor_share | 0.31 |
| 37/42 | 议题解决 — 88% 的议题已关闭 |
| 26.2/30 | PR 接受 — 已裁定的 PR 中 1,216/1,392 已合并 |
| 0/13 | Newcomer PR acceptance — 30 天内没有首次贡献者的 PR 得到裁决 |
| 15/15 | OpenSSF Scorecard:Code-Review — all changesets reviewed |
| merged_prs | 1,216 |
| open_issues | 76 |
| closed_issues | 560 |
| prs_merged_7d | — |
| prs_decided_7d | — |
| prs_merged_30d | — |
| prs_decided_30d | — |
| issue_closed_ratio | 0.881 |
| closed_unmerged_prs | 176 |
| first_time_authors_30d | — |
| first_time_prs_merged_30d | — |
| first_time_prs_decided_30d | — |
| 30/30 | 所有权背书 — 组织持有 |
| 0/20 | 已验证域名 — 未读取该组织的域名验证状态 |
| 25/25 | 所有者影响力 — intel 有 5,420 位关注者 |
| 25/25 | 既往记录 — 1,344 个公开仓库,账户约 10 年 |
| followers | 5,420 |
| owner_type | Organization |
| is_verified | — |
| owner_login | intel |
| public_repos | 1,344 |
| account_age_days | 3,773 |
基础的工程与文档实践是否到位?
| 24/24 | CI 工作流 — 5 个工作流 |
| 24/24 | 存在测试 |
| 16/16 | Linter 配置 |
| 9.6/9.6 | Pre-commit 钩子 |
| 0/6.4 | .editorconfig |
| 18/20 | OpenSSF Scorecard:CI-Tests — 29 out of 30 merged PRs checked by a CI test -- score normalized to 9 |
| has_ci | 是 |
| has_tests | 是 |
| has_editorconfig | 否 |
| has_linter_config | 是 |
| has_precommit_config | 是 |
| 30/30 | README |
| 25/25 | 文档目录 |
| 0/15 | 文档 / 主页站点 |
| 10/10 | 仓库描述 |
| 10/10 | 主题标签 — 13 个主题标签 |
| 10/10 | Wiki |
| topics | int4, quantization, rounding, transformers, vllm, mxfp4, nvfp4, gguf, sglang, llms, vlms, diffusers, omni |
| has_wiki | 是 |
| homepage | — |
| docs_site | — |
| has_readme | 是 |
| has_docs_dir | 是 |
| has_description | 是 |
可见的安全与供应链实践是否稳固,且不存在未解决的高风险司法辖区暴露?
| 7.5/7.5 | Binary-Artifacts — no binaries found in the repo |
| 2.2/7.5 | Branch-Protection — branch protection is not maximal on development and all release branches |
| 2.2/2.5 | CI-Tests — 29 out of 30 merged PRs checked by a CI test -- score normalized to 9 |
| 0/2.5 | CII-Best-Practices — no effort to earn an OpenSSF best practices badge detected |
| 7.5/7.5 | Code-Review — all changesets reviewed |
| 1.5/2.5 | Contributors — project has 2 contributing companies or organizations -- score normalized to 6 |
| 10/10 | Dangerous-Workflow — no dangerous workflow patterns detected |
| 7.5/7.5 | Dependency-Update-Tool — update tool detected |
| 0/5 | Fuzzing — project is not fuzzed |
| 2.5/2.5 | 许可证 — license file detected |
| 7.5/7.5 | Maintained — 30 commit(s) and 0 issue activity found in the last 90 days -- score normalized to 10 |
| 0/5 | Packaging — 无数据 |
| 0/5 | Pinned-Dependencies — dependency not pinned by hash detected -- score normalized to 0 |
| 4.5/5 | SAST — SAST tool detected but not run on all commits |
| 5/5 | Security-Policy — security policy file detected |
| 0/7.5 | Signed-Releases — 无数据 |
| 6.8/7.5 | Token-Permissions — detected GitHub workflow tokens with excessive permissions |
| 0/7.5 | Vulnerabilities — 28 existing vulnerabilities detected |
| source | openssf_scorecard |
| checks_evaluated | 16 |
| scorecard_version | v5.5.0 |
| checks_inconclusive | 2 |
| scorecard_aggregate | 7 |
该仓库在多大程度上具备与 AI 编码代理协同开发与维护的条件?权重刻意设小(4%):代理工具链是一项真实的维护信号,但完全不具备的仓库仍可达到 100/100。
| 45/45 | 代理指令 — .github/copilot-instructions.md, AGENTS.md, CLAUDE.md |
| 0/15 | 机器可读文档(llms.txt) |
| 0/40 | 可读的提交历史 — 无数据 |
| has_llms_txt | 否 |
| llms_txt_url | — |
| legible_history_share | — |
| agent_instruction_files | .github/copilot-instructions.md, AGENTS.md, CLAUDE.md |
| agent_instruction_max_bytes | 2,729 |
| 0/18 | 一条命令的引导启动 |
| 22/22 | 自动化测试 |
| 11/11 | Lint / 格式化配置 |
| 0/11 | 静态类型检查 |
| 10/10 | 可复现环境 — Dockerfile |
| 0/10 | 已体现的代理实践 — 无数据 |
| 0/8 | 自动化维护 — 无数据 |
| 0/10 | OpenSSF Scorecard:Pinned-Dependencies — dependency not pinned by hash detected -- score normalized to 0 |
| has_nix | 否 |
| has_tests | 是 |
| lockfiles | — |
| has_dockerfile | 是 |
| typed_language | 否 |
| bootstrap_files | — |
| has_devcontainer | 否 |
| has_linter_config | 是 |
| typecheck_configs | — |
| agent_commit_share | — |
| toolchain_manifests | — |
| dependency_bot_commit_share | — |
| 0/45 | 可类型检查的代码 — Python,未配置类型检查 |
| 52.6/55 | 可控的文件大小 — 采样的 581 个源文件中有 25 个超过 60KB |
| primary_language | Python |
| largest_source_bytes | 1,930,927 |
| source_files_sampled | 581 |
| oversized_source_files | 25 |
来自开源项目 OpenSSF Scorecard 的独立、工具无关的安全评估。每项检查奖励的是安全实践本身,而非特定供应商的工具。Scorecard 无法判定的检查项标记为 不适用,并从安全评分中剔除(绝不按零分计)。
| 10 | Binary-Artifacts | no binaries found in the repo |
| 3 | Branch-Protection | branch protection is not maximal on development and all release branches |
| 9 | CI-Tests | 29 out of 30 merged PRs checked by a CI test -- score normalized to 9 |
| 0 | CII-Best-Practices | no effort to earn an OpenSSF best practices badge detected |
| 10 | Code-Review | all changesets reviewed |
| 6 | Contributors | project has 2 contributing companies or organizations -- score normalized to 6 |
| 10 | Dangerous-Workflow | no dangerous workflow patterns detected |
| 10 | Dependency-Update-Tool | update tool detected |
| 0 | Fuzzing | project is not fuzzed |
| 10 | License | license file detected |
| 10 | Maintained | 30 commit(s) and 0 issue activity found in the last 90 days -- score normalized to 10 |
| 不适用 | Packaging | packaging workflow not detected |
| 0 | Pinned-Dependencies | dependency not pinned by hash detected -- score normalized to 0 |
| 9 | SAST | SAST tool detected but not run on all commits |
| 10 | Security-Policy | security policy file detected |
| 不适用 | Signed-Releases | no releases found |
| 9 | Token-Permissions | detected GitHub workflow tokens with excessive permissions |
| 0 | Vulnerabilities | 28 existing vulnerabilities detected |
来自 GitHub 依赖图的完整解析依赖集合:0 个直接依赖与 42 个间接(传递)软件包。仓库提交锁文件时,传递闭包才是完整的。
| 注册表 | 软件包 | 版本 | 关系 |
|---|---|---|---|
| PyPI | accelerate | — | 间接 |
| PyPI | addict | — | 间接 |
| PyPI | auto-round-lib | — | 间接 |
| PyPI | bitsandbytes | — | 间接 |
| PyPI | clip | — | 间接 |
| PyPI | compressed-tensors | — | 间接 |
| PyPI | datasets | — | 间接 |
| PyPI | diffusers | — | 间接 |
| PyPI | einops | — | 间接 |
| PyPI | flash-attn | — | 间接 |
| PyPI | gguf | — | 间接 |
| PyPI | gptqmodel | — | 间接 |
| PyPI | image-reward | — | 间接 |
| PyPI | llmcompressor | — | 间接 |
| PyPI | lm-eval | — | 间接 |
| PyPI | modelscope | — | 间接 |
| PyPI | neural-compressor-pt | — | 间接 |
| PyPI | numba | — | 间接 |
| PyPI | numpy | — | 间接 |
| PyPI | optimum | — | 间接 |
| PyPI | packaging | — | 间接 |
| PyPI | pandas | — | 间接 |
| PyPI | parameterized | — | 间接 |
| PyPI | pillow | — | 间接 |
| PyPI | protobuf | — | 间接 |
| PyPI | py-cpuinfo | — | 间接 |
| PyPI | pydantic | — | 间接 |
| PyPI | pytest | — | 间接 |
| PyPI | ray | — | 间接 |
| PyPI | sentencepiece | — | 间接 |
| PyPI | setuptools | — | 间接 |
| PyPI | sglang | — | 间接 |
| PyPI | tbb | — | 间接 |
| PyPI | tiktoken | — | 间接 |
| PyPI | timm | — | 间接 |
| PyPI | torch | — | 间接 |
| PyPI | torchvision | — | 间接 |
| PyPI | tqdm | — | 间接 |
| PyPI | transformers | — | 间接 |
| PyPI | triton | — | 间接 |
| PyPI | vllm | — | 间接 |
| PyPI | xformers | — | 间接 |
发现这份报告有不准确之处,或有想法要分享?错误的测量、未识别的工具、建议、疑问——都欢迎提出。每条消息都会被阅读并得到回复。