遵循以下最佳实践的项目将能够自愿的自我认证,并显示他们已经实现了核心基础设施计划(OpenSSF)徽章。 显示详细资料
[](https://www.bestpractices.dev/projects/1486)
<a href="https://www.bestpractices.dev/projects/1486"><img src="https://www.bestpractices.dev/projects/1486/badge"></a>
Computation using data flow graphs for scalable machine learning
Contributions howto: https://github.com/tensorflow/tensorflow/blob/master/CONTRIBUTING.md includes sections on Python/C++ style used, review standards and process.
Contributors must sign CLA: https://github.com/tensorflow/tensorflow/blob/master/CONTRIBUTING.md#contributor-license-agreements
TensorFlow is governed by the TensorFlow team, but this is not clearly and publicly documented.
In standard location: https://github.com/tensorflow/tensorflow/blob/master/CODE_OF_CONDUCT.md
See governance above. No public documentation of roles.
There is a backup person with credentials and working knowledge for all critical infrastructure. Not sure what URL to provide here -- the github teams page cannot be accessed by non-members, and internal Google pages cannot be seen either.
警告:需要URL,但找不到URL。
Truck-Factor helpfully has the information for tensorflow on the site. https://github.com/aserg-ufmg/Truck-Factor
We have a roadmap (https://www.tensorflow.org/about/roadmap), but it's too infrequently updated to count.
I addition to the documentation, we have published several papers on the architecture: https://www.tensorflow.org/about/bib.
We need to provide a security page (see also reporting above).
We have a getting started section here: https://www.tensorflow.org/get_started/get_started
We accept issues and PRs for documentation fixes and treat them the same as reports of defects to the software itself.
We haven't actively worked to achieve any badges etc., but if we did, we would also post the result (including this badge). These would likely appear on https://github.com/tensorflow/tensorflow
TensorFlow is mostly exposed as an API, with reduced accessibility concerns. However, we have tried to make sure that our website follows accessibility standards, and that UI components (TensorBoard) are accessible.
TensorFlow itself does not deal with text directly.
No passwords saved.
Where the upgrade is hard (0.12 -> 1.0) we have provided scripts to help users migrate, and we plan to do the same in the future.
Using github issues: https://github.com/tensorflow/tensorflow/issues
We have no documented reporting process in place.
See reporting earlier.
See https://github.com/tensorflow/tensorflow/blob/master/CONTRIBUTING.md#c-coding-style
Using pylint and cpplint.
Bazel builds are hermetic (and must be for distributed builds), and DO NOT respect environment variables unless those are explicitly declared to the build system (with --action_env=... or in the bazel.rc).
You can make a debug build, and if you do, the generated pip package will come with debugging symbols.
bazel does not automatically recurse into subdirectories.
bazel builds are hermetic and repeatable.
We publish pip packages to pypi.
Installation via pip uses Python's site-packages directories.
We provide docker containers equivalent to our CI setup.
See the explicit dependency declaration in our bazel workspace: https://github.com/tensorflow/tensorflow/blob/master/tensorflow/workspace.bzl
We have no process to regularly screen our dependencies.
We pin components to explicit versions. Updating requires rebuilding.
警告:需要更长的理由。
Automated tests are run on every commit to master.
Every bugfix should come with a test as described in our contributions policy: https://github.com/tensorflow/tensorflow/blob/master/CONTRIBUTING.md#general-guidelines-and-philosophy-for-contribution
Our coverage is >90%.
See contribution guidelines: https://github.com/tensorflow/tensorflow/blob/master/CONTRIBUTING.md#general-guidelines-and-philosophy-for-contribution
Our contributions guidelines state that "Include unit tests when you contribute new features, as they help to a) prove that your code works correctly, b) guard against future breaking changes to lower the maintenance cost. Bug fixes also generally require unit tests, because the presence of bugs usually indicates insufficient test coverage."
(https://github.com/tensorflow/tensorflow/blob/master/CONTRIBUTING.md#general-guidelines-and-philosophy-for-contribution)
There are likely optional warnings we could enable but don't.
All our analysis tools are general purpose. However, AddressSanitizer does have "rules to look for common vulnerabilities in the analyzed language" (buffer overruns in particular).
We do not use a fuzzer, but we do use ASAN, MSAN, and TSAN (see static analysis above).
后退