Open belldandyxtq opened 4 years ago
How about being more clear with
本函数是 optuna.study.load_study 的一个别名
Same here https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/core.po#L158 https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/core.po#L172 https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/core.po#L192
I am not sure about this one
maybe
可重复叠加
or 可叠加
will be more clear to me
I think this is confusing
该选项默认情况下是禁用的
can mean in default case we cannot use this option
How about this
指定用于记录日志的文件名。如未指定,则不记录日志。
How about
为study设置易懂的名称以区别于其他study
Can you please split this message and create different item for each one? As the message is misleading. Only create-study
will create new study
See https://github.com/optuna/optuna/pull/1565
can you translate a little more like:
Optuna 将会启动一个 HTTP 服务器并架设一个dashboard
Some typo and nits:
允许从所给的地址发起的 websocket 请求。
https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/distributions.po#L137
https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/distributions.po#L125
https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/distributions.po#L163
step == 步进
, can we make them consistent?
https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/distributions.po#L153-L158
Just curious, I do not see this part in the v1.4.0
doc.
https://optuna.readthedocs.io/en/v1.4.0/reference/distributions.html#optuna.distributions.IntUniformDistribution
I do not think the translation is correct.
It should be 所给分布的JSON字符串
参数重要性以字典形式返回,其中键是参数名称,值是其重要性。重要性由浮点数表示,一个字典中所有的重要性之和为1.0。该浮点数的值越高,其对应的参数也就越重要。这个返回的字典属于 collections.OrderedDict 类,其中的元素会按照其重要性的大小降序排列。
I think you missed punctuation at the end of each line. Please be careful. https://chinese-doc-optuna.onrender.com/reference/importance.html#optuna.importance.get_param_importances
I think it means
那么所有出现在全部已完成 trial 的参数会被评估。
也就是有所trial中参数集合的交集
for example: trial A: {a, b, c, d} trial B: {b, c, d} trial C: {a, b, d}
With params is None
, we get {b, d}
I notice that we have docs from versions higher than v1.4.0
in the doc of v1.4.0
https://chinese-doc-optuna.onrender.com/reference/importance.html#optuna.importance.MeanDecreaseImpurityImportanceEvaluator
https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/integration.po#L38-L39 https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/integration.po#L35-L37 https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/integration.po#L82-L84 The links are broken https://chinese-doc-optuna.onrender.com/reference/integration.html#optuna.integration.ChainerPruningExtension
https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/integration.po#L138-L142
Not translated 所给ChainerMN communicator
I am not pretty sure what this line means exactly without looking at the English version. How about this:
采用的独立采样而不是 CMA-ES 算法采样直到所指定的trials个数完成。
https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L328 https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L365 https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L392 https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L422 https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L712 https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L737 https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L835 https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L859 https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L881
I think the word 清除
gives a wrong idea to some readers.
For example: 用于清除 FastAI 中无望 trial 的回调函数。
, The sentence can be read as Used to remove the callback function that is used to prune trial in FastAI
, which is totally wrong.
How about we say it more clearly, like 注册FastAI 中用于剪枝 trial 的回调函数。
https://github.com/chenghuzi/doc-optuna-chinese/blob/d63596032f5f027bf68ac12336e626a6e03fb6ee/source/locale/zh/LC_MESSAGES/reference/integration.po#L491
How about 本函数接受所有lightgbm.train() 的参数
https://chinese-doc-optuna.onrender.com/reference/integration.html#optuna.integration.lightgbm.LightGBMTuner https://chinese-doc-optuna.onrender.com/reference/integration.html#optuna.integration.lightgbm.LightGBMTunerCV These part are not in the v1.4.0 documentation, but v1.5.0
https://github.com/chenghuzi/doc-optuna-chinese/blob/441ac8fcba05a5928093c596717328e53ab1f4c8/source/locale/zh/LC_MESSAGES/reference/core.po#L65
Missing period at the end