Component Translated Unfinished Unfinished words Unfinished characters Untranslated Checks Suggestions Comments
Framework 26% 20,202 327,120 3,024,555 14,304 4,025
Glossary Flower Docs

Summary

Project website flower.ai
Project maintainers2 User avatar charlesbvllUser avatar tanertopal
04/07/2026

String statistics

Strings percent Hosted strings Words percent Hosted words Characters percent Hosted characters
Total 27,425 441,635 4,009,175
Source 5,485 88,327 801,835
Translated 26% 7,223 25% 114,515 24% 984,620
Needs editing 21% 5,898 11% 50,809 16% 674,618
Read-only 20% 5,485 20% 88,327 20% 801,835
Failing checks 14% 4,025 8% 39,416 14% 590,667
Strings with suggestions 0% 0 0% 0 0% 0
Untranslated strings 52% 14,304 62% 276,311 58% 2,349,937
Last 12 months Previous

Activity in last 12 months

2,283 changes 702 changes 5,420 changes 1,086 changes 3,443 changes 918 changes 45 changes 15,600 changes 44 changes 4,528 changes 5,658 changes 7,237 changes 90 changes 702 changes 3,408 changes 350 changes 1,146 changes 710 changes 1,366 changes 73 changes 1,136 changes 0 changes 2,096 changes 5,181 changes
AprMayJunJulAugSepOctNovDecJanFebMar

Quick numbers

441 k
Hosted words
27,425
Hosted strings
26%
Translated
11
Contributors
2
Components
5
Languages
10
Translations
0
Screenshots
and previous 30 days

Trends of last 30 days

+5%
Hosted words
+100%
+4%
Hosted strings
+100%
−1%
Translated
+28%
—
Contributors
+100%
User avatar charlesbvll

Resource updated
The “framework/docs/locales/zh_Hans/LC_MESSAGES/framework-docs.po” file was changed. 04/07/2026
User avatar charlesbvll

String added in the repository
We now have 200 partitions, each holding 45 training and 5 validation examples. Given that the number of training examples on each client is quite small, we should probably train the model a bit longer, so we configure the clients to perform 3 local training epochs. We should also adjust the fraction of clients selected for training during each round (we don't want all 200 clients participating in every round), so we add ``fraction-train = 0.025`` and adjust ``fraction-evaluate`` to ``0.05``, which means that only 2.5% of available clients will be selected for training each round (so 5 clients) and 5% of them for evaluation (so 10 clients). We can add and adjust values in the ``pyproject.toml`` for ease of experimentation:
现在我们有 1000 个分区，每个分区有 45 个训练数据和 5 个验证数据。鉴于每个客户端上的训练示例数量较少，我们可能需要对模型进行更长时间的训练，因此我们将客户端配置为执行 3 个本地训练遍历。我们还应该调整每轮训练中被选中的客户端的比例（我们不希望每轮训练都有 1000 个客户端参与），因此我们将 ``fraction_fit`` 调整为 ``0.05``，这意味着每轮训练只选中 5%的可用客户端（即 50 个客户端）：
04/07/2026
User avatar charlesbvll

String added in the repository
For more details on the Simulation Runtime and its configuration, check out the :doc:`Simulation Runtime documentation <how-to-run-simulations>`.
 
04/07/2026
User avatar charlesbvll

String added in the repository
As a last step in this tutorial, let's see how we can use Flower to experiment with a large number of clients. The most straightforward way to do this is by overriding the default Simulation Runtime configuration via the ``--federation-config`` flag:
作为本笔记的最后一步，让我们看看如何使用 Flower 对大量客户端进行实验。
04/07/2026
User avatar charlesbvll

String added in the repository
When we execute ``flwr run`` against the default local connection configuration (for example, the one with ``address = ":local:"``), Flower submits the run to the managed local SuperLink. By default, the local SuperLink will configure the simulation runtime to use 10 clients. Each will run an instance of the ``ClientApp`` we defined earlier.
 
04/07/2026
User avatar charlesbvll

String added in the repository
This submits the run to the managed local SuperLink for the ``[superlink.local]`` profile (identified in the Flower configuration with ``address = ":local:"``), which then executes the federated learning simulation with 10 clients, or SuperNodes, using the Flower Simulation Runtime. Plain ``flwr run .`` would submit the run, print the run ID, and return without streaming logs. For the full local workflow, see :doc:`how-to-run-flower-locally`. If you run into SQL database errors during local simulations, see :ref:`FAQ <faq-local-superlink-db-error>`.
 
04/07/2026
User avatar charlesbvll

String added in the repository
Let's build a federated learning system using `TFLite <https://ai.google.dev/edge/litert/android>`_ and Flower on Android!
让我们在 Android 上使用 TFLite 和 Flower 构建一个联邦学习系统！
04/07/2026
User avatar charlesbvll

String added in the repository
When you use a local simulation profile with ``address = ":local:"`` or ``address = ":local-in-memory:"``, Flower CLI commands that communicate with SuperLink use the Control API. Flower starts a managed local SuperLink automatically when needed and reuses it across ``flwr run``, ``flwr list``, ``flwr log``, and ``flwr stop``. See :doc:`how-to-run-flower-locally` for the full local workflow, background process behavior, and runtime file locations.
 
04/07/2026
User avatar charlesbvll

String added in the repository
Use ``address = ":local-in-memory:"`` when the managed local SuperLink runs on a filesystem where SQLite performs poorly, such as some network drives or HPC storage
 
04/07/2026
User avatar charlesbvll

String added in the repository
Use the basic configuration for most local development and testing scenarios, especially when you want to keep a history of your runs and logs on disk.
 
04/07/2026
Browse all project changes