This is an automated email from the ASF dual-hosted git repository.
github-bot pushed a commit to branch asf-site
in repository https://gitbox.apache.org/repos/asf/dolphinscheduler-website.git
The following commit(s) were added to refs/heads/asf-site by this push:
new efdd36e1b Automated deployment:
4d97cc3dab28602c236aadcffb847cfa90bd99ad
efdd36e1b is described below
commit efdd36e1bcf2a4ed2236ae14cc8c953d3eed985f
Author: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
AuthorDate: Fri Jun 10 05:06:02 2022 +0000
Automated deployment: 4d97cc3dab28602c236aadcffb847cfa90bd99ad
---
en-us/docs/dev/user_doc/guide/task/mlflow.html | 96 +++++++++++-----------
en-us/docs/dev/user_doc/guide/task/mlflow.json | 2 +-
img/tasks/demo/mlflow-automl.png | Bin 28971 -> 32372 bytes
img/tasks/demo/mlflow-basic-algorithm.png | Bin 32819 -> 33503 bytes
img/tasks/demo/mlflow-custom-project-template.png | Bin 66296 -> 0 bytes
img/tasks/demo/mlflow-custom-project.png | Bin 84087 -> 35993 bytes
img/tasks/demo/mlflow-models-docker-compose.png | Bin 0 -> 24671 bytes
img/tasks/demo/mlflow-models-docker.png | Bin 20543 -> 17866 bytes
img/tasks/demo/mlflow-models-mlflow.png | Bin 20460 -> 17859 bytes
zh-cn/docs/dev/user_doc/guide/task/mlflow.html | 75 ++++++++---------
zh-cn/docs/dev/user_doc/guide/task/mlflow.json | 2 +-
11 files changed, 84 insertions(+), 91 deletions(-)
diff --git a/en-us/docs/dev/user_doc/guide/task/mlflow.html
b/en-us/docs/dev/user_doc/guide/task/mlflow.html
index c624e07f5..aaf7db72a 100644
--- a/en-us/docs/dev/user_doc/guide/task/mlflow.html
+++ b/en-us/docs/dev/user_doc/guide/task/mlflow.html
@@ -14,13 +14,13 @@
<h2>Overview</h2>
<p><a href="https://mlflow.org";>MLflow</a> is an excellent open source
platform to manage the ML lifecycle, including experimentation,
reproducibility, deployment, and a central model registry.</p>
-<p>MLflow task plugin used to execute MLflow tasks，Currently contains Mlflow
Projects and MLflow Models.（Model Registry will soon be rewarded for
support）</p>
+<p>MLflow task plugin used to execute MLflow tasks，Currently contains MLflow
Projects and MLflow Models. (Model Registry will soon be rewarded for
support)</p>
<ul>
-<li>Mlflow Projects: Package data science code in a format to reproduce runs
on any platform.</li>
+<li>MLflow Projects: Package data science code in a format to reproduce runs
on any platform.</li>
<li>MLflow Models: Deploy machine learning models in diverse serving
environments.</li>
<li>Model Registry: Store, annotate, discover, and manage models in a central
repository.</li>
</ul>
-<p>The Mlflow plugin currently supports and will support the following:</p>
+<p>The MLflow plugin currently supports and will support the following:</p>
<ul>
<li>[x] MLflow Projects
<ul>
@@ -33,10 +33,10 @@ reproducibility, deployment, and a central model
registry.</p>
<ul>
<li>[x] MLFLOW: Use <code>MLflow models serve</code> to deploy a model
service</li>
<li>[x] Docker: Run the container after packaging the docker image</li>
-<li>[ ] Docker Compose: Use docker compose to run the container, Will replace
the docker run above</li>
+<li>[x] Docker Compose: Use docker compose to run the container, it will
replace the docker run above</li>
<li>[ ] Seldon core: Use Selcon core to deploy model to k8s cluster</li>
<li>[ ] k8s: Deploy containers directly to K8S</li>
-<li>[ ] mlflow deployments: Built-in deployment modules, such as built-in
deployment to SageMaker, etc</li>
+<li>[ ] MLflow deployments: Built-in deployment modules, such as built-in
deployment to SageMaker, etc</li>
</ul>
</li>
<li>[ ] Model Registry
@@ -52,7 +52,7 @@ DAG editing page.</li>
<li>Drag from the toolbar <img src="/img/tasks/icons/mlflow.png" width="15"/>
task node to canvas.</li>
</ul>
<h2>Task Example</h2>
-<p>First, introduce some general parameters of DolphinScheduler</p>
+<p>First, introduce some general parameters of DolphinScheduler:</p>
<ul>
<li><strong>Node name</strong>: The node name in a workflow definition is
unique.</li>
<li><strong>Run flag</strong>: Identifies whether this node schedules
normally, if it does not need to execute, select
@@ -71,23 +71,26 @@ email will send and the task execution will fail.</li>
<li><strong>Predecessor task</strong>: Selecting a predecessor task for the
current task, will set the selected predecessor task as
upstream of the current task.</li>
</ul>
+<p>Here are some specific parameters for the MLFlow component:</p>
+<ul>
+<li><strong>MLflow Tracking Server URI</strong>: MLflow Tracking Server URI,
default <a href="http://localhost:5000";>http://localhost:5000</a>.</li>
+<li><strong>Experiment Name</strong>: Create the experiment where the task is
running, if the experiment does not exist. If the name is empty, it is set to
<code>Default</code>, the same as MLflow.</li>
+</ul>
<h3>MLflow Projects</h3>
<h4>BasicAlgorithm</h4>
<p><img src="/img/tasks/demo/mlflow-basic-algorithm.png"
alt="mlflow-conda-env"></p>
<p><strong>Task Parameter</strong></p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server uri, default <a
href="http://localhost:5000";>http://localhost:5000</a>.</li>
-<li><strong>experiment name</strong> ：Create the experiment where the task is
running, if the experiment does not exist. If the name is empty, it is set to
<code>Default</code>, the same as MLflow.</li>
-<li><strong>register model</strong> ：Register the model or not. If register is
selected, the following parameters are expanded.
+<li><strong>Register Model</strong>: Register the model or not. If register is
selected, the following parameters are expanded.
<ul>
-<li><strong>model name</strong> : The registered model name is added to the
original model version and registered as
+<li><strong>Model Name</strong>: The registered model name is added to the
original model version and registered as
Production.</li>
</ul>
</li>
-<li><strong>data path</strong> : The absolute path of the file or folder. Ends
with .csv for file or contain train.csv and
-test.csv for folder（In the suggested way, users should build their own test
sets for model evaluation）。</li>
-<li><strong>parameters</strong> : Parameter when initializing the
algorithm/AutoML model, which can be empty. For example
-parameters <code>"time_budget=30;estimator_list=['lgbm']"</code> for
flaml 。The convention will be passed with '; 'shards
+<li><strong>Data Path</strong>: The absolute path of the file or folder. Ends
with .csv for file or contain train.csv and
+test.csv for folder（In the suggested way, users should build their own test
sets for model evaluation）.</li>
+<li><strong>Parameters</strong>: Parameter when initializing the
algorithm/AutoML model, which can be empty. For example
+parameters <code>"time_budget=30;estimator_list=['lgbm']"</code> for
flaml 。The convention will be passed with '; ' shards
each parameter, using the name before the equal sign as the parameter name,
and using the name after the equal
sign to get the corresponding parameter value through <code>python
eval()</code>.
<ul>
@@ -97,9 +100,9 @@ sign to get the corresponding parameter value through
<code>python eval()</code>
<li><a
href="https://xgboost.readthedocs.io/en/stable/python/python_api.html#xgboost.XGBClassifier";>xgboost</a></li>
</ul>
</li>
-<li><strong>algorithm</strong> ：The selected algorithm currently supports
<code>LR</code>, <code>SVM</code>, <code>LightGBM</code> and
<code>XGboost</code> based
+<li><strong>Algorithm</strong>：The selected algorithm currently supports
<code>LR</code>, <code>SVM</code>, <code>LightGBM</code> and
<code>XGboost</code> based
on <a href="https://scikit-learn.org/";>scikit-learn</a> form.</li>
-<li><strong>Parameter search space</strong> : Parameter search space when
running the corresponding algorithm, which can be
+<li><strong>Parameter Search Space</strong>: Parameter search space when
running the corresponding algorithm, which can be
empty. For example, the parameter <code>max_depth=[5, 10];n_estimators=[100,
200]</code> for lightgbm 。The convention
will be passed with '; 'shards each parameter, using the name before the equal
sign as the parameter name,
and using the name after the equal sign to get the corresponding parameter
value through <code>python eval()</code>.</li>
@@ -108,18 +111,16 @@ and using the name after the equal sign to get the
corresponding parameter value
<p><img src="/img/tasks/demo/mlflow-automl.png" alt="mlflow-automl"></p>
<p><strong>Task Parameter</strong></p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server uri, default <a
href="http://localhost:5000";>http://localhost:5000</a>.</li>
-<li><strong>experiment name</strong> ：Create the experiment where the task is
running, if the experiment does not exist. If the name is empty, it is set to
<code>Default</code>, the same as MLflow.</li>
-<li><strong>register model</strong> ：Register the model or not. If register is
selected, the following parameters are expanded.
+<li><strong>Register Model</strong>: Register the model or not. If register is
selected, the following parameters are expanded.
<ul>
-<li><strong>model name</strong> : The registered model name is added to the
original model version and registered as
+<li><strong>model name</strong>: The registered model name is added to the
original model version and registered as
Production.</li>
</ul>
</li>
-<li><strong>data path</strong> : The absolute path of the file or folder. Ends
with .csv for file or contain train.csv and
-test.csv for folder（In the suggested way, users should build their own test
sets for model evaluation）。</li>
-<li><strong>parameters</strong> : Parameter when initializing the
algorithm/AutoML model, which can be empty. For example
-parameters <code>n_estimators=200;learning_rate=0.2</code> for flaml 。The
convention will be passed with '; 'shards
+<li><strong>Data Path</strong>: The absolute path of the file or folder. Ends
with .csv for file or contain train.csv and
+test.csv for folder(In the suggested way, users should build their own test
sets for model evaluation).</li>
+<li><strong>Parameters</strong>: Parameter when initializing the
algorithm/AutoML model, which can be empty. For example
+parameters <code>n_estimators=200;learning_rate=0.2</code> for flaml. The
convention will be passed with '; 'shards
each parameter, using the name before the equal sign as the parameter name,
and using the name after the equal
sign to get the corresponding parameter value through <code>python
eval()</code>. The detailed parameter list is as follows:
<ul>
@@ -127,58 +128,53 @@ sign to get the corresponding parameter value through
<code>python eval()</code>
<li><a
href="https://automl.github.io/auto-sklearn/master/api.html";>autosklearn</a></li>
</ul>
</li>
-<li><strong>AutoML tool</strong> : The AutoML tool used, currently
+<li><strong>AutoML tool</strong>: The AutoML tool used, currently
supports <a href="https://github.com/automl/auto-sklearn";>autosklearn</a>
-and <a href="https://github.com/microsoft/FLAML";>flaml</a></li>
+and <a href="https://github.com/microsoft/FLAML";>flaml</a>.</li>
</ul>
<h4>Custom projects</h4>
-<p><img src="/img/tasks/demo/mlflow-custom-project-template.png"
alt="mlflow-custom-project-template.png"></p>
+<p><img src="/img/tasks/demo/mlflow-custom-project.png"
alt="mlflow-custom-project.png"></p>
<p><strong>Task Parameter</strong></p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server uri, default <a
href="http://localhost:5000";>http://localhost:5000</a>.</li>
-<li><strong>experiment name</strong> ：Create the experiment where the task is
running, if the experiment does not exist. If the name is empty, it is set to
<code>Default</code>, the same as MLflow.</li>
-<li><strong>parameters</strong> : <code>--param-list</code> in <code>mlflow
run</code>. For example <code>-P learning_rate=0.2 -P colsample_bytree=0.8 -P
subsample=0.9</code></li>
-<li><strong>Repository</strong> : Repository url of MLflow Project，Support git
address and directory on worker. If it's in a subdirectory，We add
<code>#</code> to support this (same as <code>mlflow run</code>) , for example
<code>https://github.com/mlflow/mlflow#examples/xgboost/xgboost_native</code></li>
-<li><strong>Project Version</strong> : Version of the project，default
master</li>
+<li><strong>parameters</strong>: <code>--param-list</code> in <code>mlflow
run</code>. For example <code>-P learning_rate=0.2 -P colsample_bytree=0.8 -P
subsample=0.9</code>.</li>
+<li><strong>Repository</strong>: Repository url of MLflow Project，Support git
address and directory on worker. If it's in a subdirectory，We add
<code>#</code> to support this (same as <code>mlflow run</code>) , for example
<code>https://github.com/mlflow/mlflow#examples/xgboost/xgboost_native</code>.</li>
+<li><strong>Project Version</strong>: Version of the project，default
master.</li>
</ul>
-<p>You can now use this feature to run all mlFlow projects on Github (For
example <a href="https://github.com/mlflow/mlflow/tree/master/examples";>MLflow
examples</a> )了。You can also create your own machine learning library to reuse
your work, and then use DolphinScheduler to use your library with one click.</p>
-<p>The actual interface is as follows</p>
-<p><img src="/img/tasks/demo/mlflow-custom-project.png"
alt="mlflow-custom-project.png"></p>
+<p>You can now use this feature to run all MLFlow projects on Github (For
example <a href="https://github.com/mlflow/mlflow/tree/master/examples";>MLflow
examples</a> ). You can also create your own machine learning library to reuse
your work, and then use DolphinScheduler to use your library with one click.</p>
<h3>MLflow Models</h3>
-<h4>MLFLOW</h4>
-<p><img src="/img/tasks/demo/mlflow-models-mlflow.png"
alt="mlflow-models-mlflow"></p>
-<p><strong>Task Parameter</strong></p>
+<p>General Parameters:</p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server uri, default <a
href="http://localhost:5000";>http://localhost:5000</a>.</li>
-<li><strong>model-uri</strong> ：Model-uri of mlflow , support
<code>models:/<model_name>/suffix</code> format and <code>runs:/</code>
format. See <a
href="https://mlflow.org/docs/latest/tracking.html#artifact-stores";>https://mlflow.org/docs/latest/tracking.html#artifact-stores</a></li>
-<li><strong>Port</strong> ：The port to listen on</li>
+<li><strong>Model-URI</strong>: Model-URI of MLflow , support
<code>models:/<model_name>/suffix</code> format and <code>runs:/</code>
format. See <a
href="https://mlflow.org/docs/latest/tracking.html#artifact-stores";>https://mlflow.org/docs/latest/tracking.html#artifact-stores</a>.</li>
+<li><strong>Port</strong>: The port to listen on.</li>
</ul>
+<h4>MLFLOW</h4>
+<p><img src="/img/tasks/demo/mlflow-models-mlflow.png"
alt="mlflow-models-mlflow"></p>
<h4>Docker</h4>
<p><img src="/img/tasks/demo/mlflow-models-docker.png"
alt="mlflow-models-docker"></p>
-<p><strong>Task Parameter</strong></p>
+<h4>DOCKER COMPOSE</h4>
+<p><img src="/img/tasks/demo/mlflow-models-docker-compose.png"
alt="mlflow-models-docker-compose"></p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server uri, default <a
href="http://localhost:5000";>http://localhost:5000</a>.</li>
-<li><strong>model-uri</strong> ：Model-uri of mlflow , support
<code>models:/<model_name>/suffix</code> format and <code>runs:/</code>
format. See <a
href="https://mlflow.org/docs/latest/tracking.html#artifact-stores";>https://mlflow.org/docs/latest/tracking.html#artifact-stores</a></li>
-<li><strong>Port</strong> ：The port to listen on</li>
+<li><strong>Max Cpu Limit</strong>: For example <code>1.0</code> or
<code>0.5</code>, the same as docker compose.</li>
+<li><strong>Max Memory Limit</strong>: For example <code>1G</code> or
<code>500M</code>, the same as docker compose.</li>
</ul>
<h2>Environment to prepare</h2>
<h3>Conda env</h3>
<p>You need to enter the admin account to configure a conda environment
variable（Please
install <a href="https://docs.continuum.io/anaconda/install/";>anaconda</a>
-or <a
href="https://docs.conda.io/en/latest/miniconda.html#installing";>miniconda</a>
in advance )</p>
+or <a
href="https://docs.conda.io/en/latest/miniconda.html#installing";>miniconda</a>
in advance).</p>
<p><img src="/img/tasks/demo/mlflow-conda-env.png" alt="mlflow-conda-env"></p>
<p>Note During the configuration task, select the conda environment created
above. Otherwise, the program cannot find the
Conda environment.</p>
<p><img src="/img/tasks/demo/mlflow-set-conda-env.png"
alt="mlflow-set-conda-env"></p>
<h3>Start the mlflow service</h3>
-<p>Make sure you have installed MLflow, using 'PIP Install MLFlow'.</p>
-<p>Create a folder where you want to save your experiments and models and
start mlFlow service.</p>
+<p>Make sure you have installed MLflow, using 'pip install mlflow'.</p>
+<p>Create a folder where you want to save your experiments and models and
start MLflow service.</p>
<pre><code class="language-sh">mkdir mlflow
<span class="hljs-built_in">cd</span> mlflow
mlflow server -h 0.0.0.0 -p 5000 --serve-artifacts --backend-store-uri
sqlite:///mlflow.db
</code></pre>
-<p>After running, an MLflow service is started</p>
-<p>After this, you can visit the MLFlow service
(<code>http://localhost:5000</code>) page to view the experiments and
models.</p>
+<p>After running, an MLflow service is started.</p>
+<p>After this, you can visit the MLflow service
(<code>http://localhost:5000</code>) page to view the experiments and
models.</p>
<p><img src="/img/tasks/demo/mlflow-server.png" alt="mlflow-server"></p>
</div></section><footer class="footer-container"><div
class="footer-body"><div><h3>About us</h3><h4>Do you need feedback? Please
contact us through the following ways.</h4></div><div
class="contact-container"><ul><li><a
href="/en-us/community/development/subscribe.html"><img class="img-base"
src="/img/emailgray.png"/><img class="img-change"
src="/img/emailblue.png"/><p>Email List</p></a></li><li><a
href="https://twitter.com/dolphinschedule";><img class="img-base"
src="/img/twittergray.png [...]
<script src="/asset/js/react/react-with-addons.min.js"></script>
diff --git a/en-us/docs/dev/user_doc/guide/task/mlflow.json
b/en-us/docs/dev/user_doc/guide/task/mlflow.json
index e4b4ac0c9..d239abe05 100644
--- a/en-us/docs/dev/user_doc/guide/task/mlflow.json
+++ b/en-us/docs/dev/user_doc/guide/task/mlflow.json
@@ -1,6 +1,6 @@
{
"filename": "mlflow.md",
- "__html": "<h1>MLflow Node</h1>\n<h2>Overview</h2>\n<p><a
href=\"https://mlflow.org\";>MLflow</a> is an excellent open source platform to
manage the ML lifecycle, including experimentation,\nreproducibility,
deployment, and a central model registry.</p>\n<p>MLflow task plugin used to
execute MLflow tasks，Currently contains Mlflow Projects and MLflow
Models.（Model Registry will soon be rewarded for support）</p>\n<ul>\n<li>Mlflow
Projects: Package data science code in a format to reproduc [...]
+ "__html": "<h1>MLflow Node</h1>\n<h2>Overview</h2>\n<p><a
href=\"https://mlflow.org\";>MLflow</a> is an excellent open source platform to
manage the ML lifecycle, including experimentation,\nreproducibility,
deployment, and a central model registry.</p>\n<p>MLflow task plugin used to
execute MLflow tasks，Currently contains MLflow Projects and MLflow Models.
(Model Registry will soon be rewarded for support)</p>\n<ul>\n<li>MLflow
Projects: Package data science code in a format to reprodu [...]
"link": "/dist/en-us/docs/dev/user_doc/guide/task/mlflow.html",
"meta": {}
}
\ No newline at end of file
diff --git a/img/tasks/demo/mlflow-automl.png b/img/tasks/demo/mlflow-automl.png
index afdeeb7aa..66ba5c75c 100644
Binary files a/img/tasks/demo/mlflow-automl.png and
b/img/tasks/demo/mlflow-automl.png differ
diff --git a/img/tasks/demo/mlflow-basic-algorithm.png
b/img/tasks/demo/mlflow-basic-algorithm.png
index 3575ce905..bbcb488ec 100644
Binary files a/img/tasks/demo/mlflow-basic-algorithm.png and
b/img/tasks/demo/mlflow-basic-algorithm.png differ
diff --git a/img/tasks/demo/mlflow-custom-project-template.png
b/img/tasks/demo/mlflow-custom-project-template.png
deleted file mode 100644
index cbb2d6f73..000000000
Binary files a/img/tasks/demo/mlflow-custom-project-template.png and /dev/null
differ
diff --git a/img/tasks/demo/mlflow-custom-project.png
b/img/tasks/demo/mlflow-custom-project.png
index 54a8cb0ab..74c00ae1c 100644
Binary files a/img/tasks/demo/mlflow-custom-project.png and
b/img/tasks/demo/mlflow-custom-project.png differ
diff --git a/img/tasks/demo/mlflow-models-docker-compose.png
b/img/tasks/demo/mlflow-models-docker-compose.png
new file mode 100644
index 000000000..7aad9641a
Binary files /dev/null and b/img/tasks/demo/mlflow-models-docker-compose.png
differ
diff --git a/img/tasks/demo/mlflow-models-docker.png
b/img/tasks/demo/mlflow-models-docker.png
index c44665102..30587322b 100644
Binary files a/img/tasks/demo/mlflow-models-docker.png and
b/img/tasks/demo/mlflow-models-docker.png differ
diff --git a/img/tasks/demo/mlflow-models-mlflow.png
b/img/tasks/demo/mlflow-models-mlflow.png
index 677441a16..7ce41e178 100644
Binary files a/img/tasks/demo/mlflow-models-mlflow.png and
b/img/tasks/demo/mlflow-models-mlflow.png differ
diff --git a/zh-cn/docs/dev/user_doc/guide/task/mlflow.html
b/zh-cn/docs/dev/user_doc/guide/task/mlflow.html
index 5253b430f..c358dde53 100644
--- a/zh-cn/docs/dev/user_doc/guide/task/mlflow.html
+++ b/zh-cn/docs/dev/user_doc/guide/task/mlflow.html
@@ -13,29 +13,29 @@
<div id="root"><div class="md2html docs-page" data-reactroot=""><header
class="header-container header-container-dark"><div class="banner-tips"><div>🤔
有关于 Apache DolphinScheduler 的疑问，加入 Slack 频道来讨论他们 <a class="link-tips"
href="https://s.apache.org/dolphinscheduler-slack";>join #dolphinscheduler
channel</a>! 🌟</div></div><div class="header-body"><span class="mobile-menu-btn
mobile-menu-btn-dark"></span><a href="/zh-cn/index.html"><img class="logo"
src="/img/hlogo_white.svg"/></a><div cla [...]
<h2>综述</h2>
<p><a href="https://mlflow.org";>MLflow</a> 是一个MLops领域一个优秀的开源项目，
用于管理机器学习的生命周期，包括实验、可再现性、部署和中心模型注册。</p>
-<p>MLflow 组件用于执行 MLflow 任务，目前包含Mlflow Projects, 和MLflow Models。（Model
Registry将在不就的将来支持）</p>
+<p>MLflow 组件用于执行 MLflow 任务，目前包含Mlflow Projects, 和MLflow Models。（Model
Registry将在不就的将来支持）。</p>
<ul>
-<li>Mlflow Projects: 将代码打包，并可以运行到任务的平台上。</li>
+<li>MLflow Projects: 将代码打包，并可以运行到任务的平台上。</li>
<li>MLflow Models: 在不同的服务环境中部署机器学习模型。</li>
-<li>Model Registry: 在一个中央存储库中存储、注释、发现和管理模型 (你也可以在你的mlflow project
里面自行注册模型)。</li>
+<li>Model Registry: 在一个中央存储库中存储、注释、发现和管理模型 (你也可以在你的MLflow project
里面自行注册模型)。</li>
</ul>
<p>目前 Mlflow 组件支持的和即将支持的内容如下中：</p>
<ul>
<li>[x] MLflow Projects
<ul>
-<li>[x] BasicAlgorithm: 基础算法，包含LogisticRegression, svm, lightgbm, xgboost。</li>
-<li>[x] AutoML: AutoML工具，包含autosklean, flaml。</li>
+<li>[x] BasicAlgorithm: 基础算法，包含LogisticRegression, svm, lightgbm, xgboost</li>
+<li>[x] AutoML: AutoML工具，包含autosklean, flaml</li>
<li>[x] Custom projects: 支持运行自己的MLflow Projects项目</li>
</ul>
</li>
<li>[ ] MLflow Models
<ul>
-<li>[x] MLFLOW: 直接使用 <code>MLflow models serve</code> 部署模型</li>
-<li>[x] Docker: 打包 DOCKER 镜像后部署模型</li>
-<li>[ ] Docker Compose: 使用Docker Compose 部署模型，将会取代上面的Docker部署</li>
-<li>[ ] Seldon core: 构建完镜像后，使用Seldon Core 部署到k8s集群上, 可以使用Seldon
Core的生成模型管理能力</li>
-<li>[ ] k8s: 构建完镜像后， 部署到k8s集群上</li>
-<li>[ ] mlflow deployments: 内置的允许MLflow 部署模块, 如内置的部署到Sagemaker等</li>
+<li>[x] MLFLOW: 直接使用 <code>mlflow models serve</code> 部署模型。</li>
+<li>[x] Docker: 打包 DOCKER 镜像后部署模型。</li>
+<li>[x] Docker Compose: 使用Docker Compose 部署模型，将会取代上面的Docker部署。</li>
+<li>[ ] Seldon core: 构建完镜像后，使用Seldon Core 部署到k8s集群上, 可以使用Seldon
Core的生成模型管理能力。</li>
+<li>[ ] k8s: 构建完镜像后， 部署到k8s集群上。</li>
+<li>[ ] MLflow deployments: 内置的允许MLflow 部署模块, 如内置的部署到Sagemaker等。</li>
</ul>
</li>
<li>[ ] Model Registry
@@ -64,13 +64,16 @@
<li><strong>超时告警</strong>
：勾选超时告警、超时失败，当任务超过"超时时长"后，会发送告警邮件并且任务执行失败。</li>
<li><strong>前置任务</strong> ：选择当前任务的前置任务，会将被选择的前置任务设置为当前任务的上游。</li>
</ul>
+<p>以下是一些MLflow 组件的常用参数</p>
+<ul>
+<li><strong>MLflow Tracking Server URI</strong> ：MLflow Tracking Server 的连接,
默认 <a href="http://localhost:5000";>http://localhost:5000</a>。</li>
+<li><strong>实验名称</strong>
：任务运行时所在的实验，若实验不存在，则创建。若实验名称为空，则设置为<code>Default</code>, 与 MLflow 一样。</li>
+</ul>
<h3>MLflow Projects</h3>
<h4>BasicAlgorithm</h4>
<p><img src="/img/tasks/demo/mlflow-basic-algorithm.png"
alt="mlflow-conda-env"></p>
<p><strong>任务参数</strong></p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server 的连接, 默认 <a
href="http://localhost:5000";>http://localhost:5000</a>。</li>
-<li><strong>实验名称</strong>
：任务运行时所在的实验，若实验不存在，则创建。若实验名称为空，则设置为<code>Default</code>, 与 MLflow 一样。</li>
<li><strong>注册模型</strong> ：是否注册模型，若选择注册，则会展开以下参数。
<ul>
<li><strong>注册的模型名称</strong> : 注册的模型名称，会在原来的基础上加上一个模型版本，并注册为Production。</li>
@@ -85,15 +88,13 @@
<li><a
href="https://xgboost.readthedocs.io/en/stable/python/python_api.html#xgboost.XGBClassifier";>xgboost</a></li>
</ul>
</li>
-<li><strong>算法</strong> ：选择的算法，目前基于 <a
href="https://scikit-learn.org/";>scikit-learn</a> 形式支持 <code>lr</code>,
<code>svm</code>, <code>lightgbm</code>, <code>xgboost</code>.</li>
+<li><strong>算法</strong> ：选择的算法，目前基于 <a
href="https://scikit-learn.org/";>scikit-learn</a> 形式支持 <code>lr</code>,
<code>svm</code>, <code>lightgbm</code>, <code>xgboost</code>。</li>
<li><strong>参数搜索空间</strong> : 运行对应算法的参数搜索空间, 可为空。如针对lightgbm 的
<code>max_depth=[5, 10];n_estimators=[100, 200]</code>
则会进行对应搜索。约定传入后会以;切分各个参数，等号前的名字作为参数名，等号后的名字将以python eval执行得到对应的参数值</li>
</ul>
<h4>AutoML</h4>
<p><img src="/img/tasks/demo/mlflow-automl.png" alt="mlflow-automl"></p>
<p><strong>任务参数</strong></p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server 的连接, 默认 <a
href="http://localhost:5000";>http://localhost:5000</a>。</li>
-<li><strong>实验名称</strong>
：任务运行时所在的实验，若实验不存在，则创建。若实验名称为空，则设置为<code>Default</code>, 与 MLflow 一样。</li>
<li><strong>注册模型</strong> ：是否注册模型，若选择注册，则会展开以下参数。
<ul>
<li><strong>注册的模型名称</strong> : 注册的模型名称，会在原来的基础上加上一个模型版本，并注册为Production。</li>
@@ -107,53 +108,49 @@
</ul>
</li>
<li><strong>AutoML工具</strong> : 使用的AutoML工具，目前支持 <a
href="https://github.com/automl/auto-sklearn";>autosklearn</a>
-, <a href="https://github.com/microsoft/FLAML";>flaml</a></li>
+, <a href="https://github.com/microsoft/FLAML";>flaml</a>。</li>
</ul>
<h4>Custom projects</h4>
-<p><img src="/img/tasks/demo/mlflow-custom-project-template.png"
alt="mlflow-custom-project-template.png"></p>
+<p><img src="/img/tasks/demo/mlflow-custom-project.png"
alt="mlflow-custom-project.png"></p>
<p><strong>任务参数</strong></p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server 的连接, 默认 <a
href="http://localhost:5000";>http://localhost:5000</a>。</li>
-<li><strong>实验名称</strong>
：任务运行时所在的实验，若实验不存在，则创建。若实验名称为空，则设置为<code>Default</code>, 与 MLflow 一样。</li>
<li><strong>参数</strong> : <code>mlflow run</code>中的 --param-list 如 <code>-P
learning_rate=0.2 -P colsample_bytree=0.8 -P subsample=0.9</code></li>
-<li><strong>运行仓库</strong> : MLflow Project的仓库地址，可以为github地址，或者worker上的目录,
如Mlflow project位于子目录，可以添加 <code>#</code> 隔开, 如
<code>https://github.com/mlflow/mlflow#examples/xgboost/xgboost_native</code></li>
+<li><strong>运行仓库</strong> : MLflow Project的仓库地址，可以为github地址，或者worker上的目录,
如MLflow project位于子目录，可以添加 <code>#</code> 隔开, 如
<code>https://github.com/mlflow/mlflow#examples/xgboost/xgboost_native</code></li>
<li><strong>项目版本</strong> : 对应项目中git版本管理中的版本，默认 master</li>
</ul>
<p>现在你可以使用这个功能来运行github上所有的MLflow Projects (如 <a
href="https://github.com/mlflow/mlflow/tree/master/examples";>MLflow
examples</a>
)了。你也可以创建自己的机器学习库，用来复用你的研究成果，以后你就可以使用DolphinScheduler来一键操作使用你的算法库。</p>
-<p>实际运行界面如下</p>
-<p><img src="/img/tasks/demo/mlflow-custom-project.png"
alt="mlflow-custom-project.png"></p>
<h3>MLflow Models</h3>
-<h4>MLFLOW</h4>
-<p><img src="/img/tasks/demo/mlflow-models-mlflow.png"
alt="mlflow-models-mlflow"></p>
-<p><strong>任务参数</strong></p>
+<p>常用参数:</p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server 的连接, 默认 <a
href="http://localhost:5000";>http://localhost:5000</a>。</li>
-<li><strong>部署模型的uri</strong> ：mlflow 服务里面模型对应的uri, 支持
<code>models:/<model_name>/suffix</code> 格式 和 <code>runs:/</code> 格式。</li>
-<li><strong>部署端口</strong> ：部署服务时的端口。</li>
+<li><strong>部署模型的URI</strong> ：MLflow 服务里面模型对应的URI, 支持
<code>models:/<model_name>/suffix</code> 格式 和 <code>runs:/</code> 格式。</li>
+<li><strong>监听端口</strong> ：部署服务时的端口。</li>
</ul>
+<h4>MLFLOW</h4>
+<p><img src="/img/tasks/demo/mlflow-models-mlflow.png"
alt="mlflow-models-mlflow"></p>
<h4>Docker</h4>
<p><img src="/img/tasks/demo/mlflow-models-docker.png"
alt="mlflow-models-docker"></p>
+<h4>DOCKER COMPOSE</h4>
+<p><img src="/img/tasks/demo/mlflow-models-docker-compose.png"
alt="mlflow-models-docker-compose"></p>
<ul>
-<li><strong>mlflow server tracking uri</strong> ：MLflow server 的连接, 默认 <a
href="http://localhost:5000";>http://localhost:5000</a>。</li>
-<li><strong>部署模型的uri</strong> ：mlflow 服务里面模型对应的uri, 支持
<code>models:/<model_name>/suffix</code> 格式 和 <code>runs:/</code> 格式。</li>
-<li><strong>部署端口</strong> ：部署服务时的端口。</li>
+<li><strong>最大CPU限制</strong> ：如 <code>1.0</code> 或者 <code>0.5</code>, 与 docker
compose 一致。</li>
+<li><strong>最大内存限制</strong> ：如 <code>1G</code> 或者 <code>500M</code>, 与 docker
compose 一致。</li>
</ul>
<h2>环境准备</h2>
<h3>conda 环境配置</h3>
<p>你需要进入admin账户配置一个conda环境变量（请提前<a
href="https://docs.continuum.io/anaconda/install/";>安装anaconda</a>
-或者<a
href="https://docs.conda.io/en/latest/miniconda.html#installing";>安装miniconda</a>
)</p>
+或者<a
href="https://docs.conda.io/en/latest/miniconda.html#installing";>安装miniconda</a>
)。</p>
<p><img src="/img/tasks/demo/mlflow-conda-env.png" alt="mlflow-conda-env"></p>
-<p>后续注意配置任务时，环境选择上面创建的conda环境，否则程序会找不到conda环境</p>
+<p>后续注意配置任务时，环境选择上面创建的conda环境，否则程序会找不到conda环境。</p>
<p><img src="/img/tasks/demo/mlflow-set-conda-env.png"
alt="mlflow-set-conda-env"></p>
-<h3>mlflow service 启动</h3>
-<p>确保你已经安装mlflow，可以使用<code>pip install mlflow</code>进行安装</p>
-<p>在你想保存实验和模型的地方建立一个文件夹，然后启动 mlflow service</p>
+<h3>MLflow service 启动</h3>
+<p>确保你已经安装MLflow，可以使用<code>pip install mlflow</code>进行安装。</p>
+<p>在你想保存实验和模型的地方建立一个文件夹，然后启动 mlflow service。</p>
<pre><code class="language-sh">mkdir mlflow
<span class="hljs-built_in">cd</span> mlflow
mlflow server -h 0.0.0.0 -p 5000 --serve-artifacts --backend-store-uri
sqlite:///mlflow.db
</code></pre>
-<p>运行后会启动一个mlflow服务</p>
-<p>可以通过访问 mlflow service (<code>http://localhost:5000</code>) 页面查看实验与模型</p>
+<p>运行后会启动一个MLflow服务。</p>
+<p>可以通过访问 MLflow service (<code>http://localhost:5000</code>) 页面查看实验与模型。</p>
<p><img src="/img/tasks/demo/mlflow-server.png" alt="mlflow-server"></p>
</div></section><footer class="footer-container"><div
class="footer-body"><div><h3>联系我们</h3><h4>有问题需要反馈？请通过以下方式联系我们。</h4></div><div
class="contact-container"><ul><li><a
href="/zh-cn/community/development/subscribe.html"><img class="img-base"
src="/img/emailgray.png"/><img class="img-change"
src="/img/emailblue.png"/><p>邮件列表</p></a></li><li><a
href="https://twitter.com/dolphinschedule";><img class="img-base"
src="/img/twittergray.png"/><img class="img-change"
src="/img/twitterblue.png"/><p [...]
<script src="/asset/js/react/react-with-addons.min.js"></script>
diff --git a/zh-cn/docs/dev/user_doc/guide/task/mlflow.json
b/zh-cn/docs/dev/user_doc/guide/task/mlflow.json
index 86e4c8c40..537e11708 100644
--- a/zh-cn/docs/dev/user_doc/guide/task/mlflow.json
+++ b/zh-cn/docs/dev/user_doc/guide/task/mlflow.json
@@ -1,6 +1,6 @@
{
"filename": "mlflow.md",
- "__html": "<h1>MLflow节点</h1>\n<h2>综述</h2>\n<p><a
href=\"https://mlflow.org\";>MLflow</a> 是一个MLops领域一个优秀的开源项目，
用于管理机器学习的生命周期，包括实验、可再现性、部署和中心模型注册。</p>\n<p>MLflow 组件用于执行 MLflow 任务，目前包含Mlflow
Projects, 和MLflow Models。（Model Registry将在不就的将来支持）</p>\n<ul>\n<li>Mlflow
Projects: 将代码打包，并可以运行到任务的平台上。</li>\n<li>MLflow Models:
在不同的服务环境中部署机器学习模型。</li>\n<li>Model Registry: 在一个中央存储库中存储、注释、发现和管理模型
(你也可以在你的mlflow project 里面自行注册模型)。</li>\n</ul>\n<p>目前 Mlflow
组件支持的和即将支持的内容如下中：</p>\n<ul>\n<li>[x] MLflow Pro [...]
+ "__html": "<h1>MLflow节点</h1>\n<h2>综述</h2>\n<p><a
href=\"https://mlflow.org\";>MLflow</a> 是一个MLops领域一个优秀的开源项目，
用于管理机器学习的生命周期，包括实验、可再现性、部署和中心模型注册。</p>\n<p>MLflow 组件用于执行 MLflow 任务，目前包含Mlflow
Projects, 和MLflow Models。（Model Registry将在不就的将来支持）。</p>\n<ul>\n<li>MLflow
Projects: 将代码打包，并可以运行到任务的平台上。</li>\n<li>MLflow Models:
在不同的服务环境中部署机器学习模型。</li>\n<li>Model Registry: 在一个中央存储库中存储、注释、发现和管理模型
(你也可以在你的MLflow project 里面自行注册模型)。</li>\n</ul>\n<p>目前 Mlflow
组件支持的和即将支持的内容如下中：</p>\n<ul>\n<li>[x] MLflow Pr [...]
"link": "/dist/zh-cn/docs/dev/user_doc/guide/task/mlflow.html",
"meta": {}
}
\ No newline at end of file
