Docs (#760)

* temp commit * tutorial * add io * temp commit * temp commit * bug fix * tutorial
keras-team · Aug 30, 2019 · 13c2850 · 13c2850
1 parent 8b1a2eb
commit 13c2850
Show file tree

Hide file tree

Showing 12 changed files with 233 additions and 13 deletions.
diff --git a/.travis.yml b/.travis.yml
@@ -6,7 +6,6 @@ jobs:
     - stage: test
       script:
         - pip install -e .[tests] --progress-bar off
-        - pip install git+git://github.com/keras-team/keras-tuner@master#egg=keras-tuner
         - pip install codacy-coverage
         - pytest tests --cov=autokeras --cov-report xml:coverage.xml
         - if ! [[ -z $CODACY_PROJECT_TOKEN ]]; then
@@ -16,12 +15,10 @@ jobs:
     -
       script:
         - pip install -e .[tests] --progress-bar off
-        - pip install git+git://github.com/keras-team/keras-tuner@master#egg=keras-tuner
         - flake8
     -
       script:
         - pip install -e .[tests] --progress-bar off
-        - pip install git+git://github.com/keras-team/keras-tuner@master#egg=keras-tuner
         - pip install mkdocs
         - pip install mkdocs-material
         - sh shell/docs.sh

diff --git a/README.md b/README.md
@@ -14,14 +14,14 @@ It is developed by <a href="http://faculty.cs.tamu.edu/xiahu/index.html" target=
 The ultimate goal of AutoML is to provide easily accessible deep learning tools to domain experts with limited data science or machine learning background.
 Auto-Keras provides functions to automatically search for architecture and hyperparameters of deep learning models.
 
-**Now we are refactoring the code on `master` branch for the next release.
-Please use the `legacy` branch if you want to checkout the 0.4 version.**
+# AutoKeras 1.0 is coming soon!
 
 ## Installation
 
 To install the package, please use the `pip` installation as follows:
 
-    pip3 install autokeras
+    pip3 install autokeras # for 0.4 version
+    pip3 install git+git://github.com/keras-team/autokeras@master#egg=autokeras # for 1.0 version
 
 **Note:** currently, Auto-Keras is only compatible with: **Python 3.6**.
 
@@ -37,6 +37,8 @@ clf.fit(x_train, y_train)
 results = clf.predict(x_test)
 ```
 
+For detailed tutorial, please check [here](https://autokeras.com/tutorial/).
+
 ## Cite this work
 
 Haifeng Jin, Qingquan Song, and Xia Hu. "Auto-keras: An efficient neural architecture search system." Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. ACM, 2019. ([Download](https://www.kdd.org/kdd2019/accepted-papers/view/auto-keras-an-efficient-neural-architecture-search-system))

diff --git a/autokeras/__init__.py b/autokeras/__init__.py
@@ -16,6 +16,7 @@
 from autokeras.hypermodel.hyperblock import LightGBMRegressorBlock
 from autokeras.hypermodel.node import ImageInput
 from autokeras.hypermodel.node import Input
+from autokeras.hypermodel.node import StructuredInput
 from autokeras.hypermodel.node import TextInput
 from autokeras.hypermodel.preprocessor import Normalization
 from autokeras.hypermodel.preprocessor import TextToIntSequence

diff --git a/autokeras/hypermodel/hyperblock.py b/autokeras/hypermodel/hyperblock.py
@@ -185,7 +185,14 @@ def build(self, hp, inputs=None):
 
 
 class GeneralBlock(HyperBlock):
-    """A general neural network block when the input type is unknown. """
+    """A general neural network block when the input type is unknown.
+
+    When the input type is unknown. The GeneralBlock would search in a large space
+    for a good model.
+
+    # Arguments
+        name: String.
+    """
 
     def build(self, hp, inputs=None):
         raise NotImplementedError
diff --git a/docs/autogen.py b/docs/autogen.py
@@ -11,7 +11,10 @@
 import six
 import autokeras
 from autokeras import auto_model
+from autokeras import task
 from autokeras.hypermodel import block
+from autokeras.hypermodel import head
+from autokeras.hypermodel import hyperblock
 
 try:
     import pathlib
@@ -114,15 +117,39 @@
                 auto_model.AutoModel.fit,
                 auto_model.AutoModel.predict,
             ]),
+        ]
+    },
+    {
+        'page': 'graph_auto_model.md',
+        'classes': [
             (auto_model.GraphAutoModel, [
                 auto_model.GraphAutoModel.fit,
                 auto_model.GraphAutoModel.predict,
             ]),
         ]
     },
     {
-        'page': 'hypermodel/block.md',
-        'all_module_classes': [block],
+        'page': 'block.md',
+        'classes': [hyperblock.ImageBlock,
+                    hyperblock.TextBlock,
+                    hyperblock.StructuredDataBlock,
+                    block.ResNetBlock,
+                    block.XceptionBlock,
+                    block.ConvBlock,
+                    block.RNNBlock,
+                    block.Merge],
+    },
+    {
+        'page': 'task.md',
+        'classes': [task.ImageClassifier,
+                    task.ImageRegressor,
+                    task.TextClassifier,
+                    task.TextRegressor],
+    },
+    {
+        'page': 'head.md',
+        'classes': [head.ClassificationHead,
+                    head.RegressionHead],
     },
     # {
     #     'page': 'hypermodel/block.md',

diff --git a/docs/mkdocs.yml b/docs/mkdocs.yml
@@ -17,11 +17,13 @@ google_analytics: ['UA-44322747-3', 'autokeras.com']
 nav:
     - Home: index.md
     - Getting Started: start.md
+    - Tutorial for 1.0: tutorial.md
     - Docker: docker.md
     - Contributing Guide: contributing.md
-    - Neural Architecture Search: nas.md
     - Documentation:
-      - auto_model: auto_model.md
-      - hypermodel:
-        - block: hypermodel/block.md
+      - Task API: task.md
+      - AutoModel: auto_model.md
+      - GraphAutoModel: graph_auto_model.md
+      - Block: block.md
+      - Head: head.md
     - About: about.md
diff --git a/docs/templates/task.md b/docs/templates/task.md
@@ -0,0 +1,13 @@
+# Task API
+
+AutoKeras support the following task APIs.
+
+{{autogenerated}}
+
+### Coming Soon:
+
+StructuredDataClassifier
+
+StructuredDataRegressor
+
+TimeSeriesForecaster
diff --git a/docs/templates/tutorial.md b/docs/templates/tutorial.md
@@ -0,0 +1,102 @@
+# AutoKeras 1.0 Tutorial
+
+In AutoKeras, there are 3 levels of APIs: task API, IO API, and functional API.
+
+## Task API
+We have designed an extremely simple interface for a series of tasks.
+The following code example shows how to do image classification with the task API.
+
+```python
+import autokeras as ak
+from keras.datasets import mnist
+
+# Prepare the data.
+(x_train, y_train), (x_test, y_test) = mnist.load_data()
+x_train = x_train.reshape(x_train.shape + (1,))
+x_test = x_test.reshape(x_test.shape + (1,))
+
+# Search and train the classifier.
+clf = ak.ImageClassifier(max_trials=100)
+clf.fit(x_train, y_train)
+y = clf.predict(x_test, y_test)
+```
+
+See the [documentation of Task API](/task) for more details.
+
+
+
+## IO API
+
+The following code example shows how to use IO API for multi-modal and multi-task scenarios using [AutoModel](/auto_model)
+
+```python
+import numpy as np
+import autokeras as ak
+from keras.datasets import mnist
+
+# Prepare the data.
+(x_train, y_classification), (x_test, y_test) = mnist.load_data()
+x_image = x_train.reshape(x_train.shape + (1,))
+x_test = x_test.reshape(x_test.shape + (1,))
+
+x_structured = np.random.rand(x_train.shape[0], 100)
+y_regression = np.random.rand(x_train.shape[0], 1)
+
+# Build model and train.
+automodel = ak.AutoModel(
+   inputs=[ak.ImageInput(),
+           ak.StructuredDataInput()],
+   outputs=[ak.RegressionHead(metrics=['mae']),
+            ak.ClassificationHead(loss='categorical_crossentropy',
+                                  metrics=['accuracy'])])
+automodel.fit([x_image, x_structured],
+              [y_regression, y_classification])
+
+```
+
+Now we support `ImageInput`, `TextInput`, and `StructuredDataInput`.
+
+## Functional API
+
+You can also define your own neural architecture with the predefined blocks and [GraphAutoModel](/graph_auto_model).
+
+```python
+import autokeras as ak
+import numpy as np
+import tensorflow as tf
+from keras.datasets import mnist
+
+# Prepare the data.
+(x_train, y_classification), (x_test, y_test) = mnist.load_data()
+x_image = x_train.reshape(x_train.shape + (1,))
+x_test = x_test.reshape(x_test.shape + (1,))
+
+x_structured = np.random.rand(x_train.shape[0], 100)
+y_regression = np.random.rand(x_train.shape[0], 1)
+
+# Build model and train.
+inputs = ak.ImageInput(shape=(28, 28, 1))
+outputs1 = ak.ResNetBlock(version='next')(inputs)
+outputs2 = ak.XceptionBlock()(inputs)
+image_outputs = ak.Merge()((outputs1, outputs2))
+
+structured_inputs = ak.StructuredInput()
+structured_outputs = ak.DenseBlock()(structured_inputs)
+merged_outputs = ak.Merge()((image_outputs, structured_outputs))
+
+classification_outputs = ak.ClassificationHead()(merged_outputs)
+regression_outputs = ak.RegressionHead()(merged_outputs)
+automodel = ak.GraphAutoModel(inputs=inputs,
+                              outputs=[regression_outputs,
+                                       classification_outputs])
+
+automodel.fit((x_image, x_structured),
+              (y_regression, y_classification),
+              trials=100,
+              epochs=200,
+              callbacks=[tf.keras.callbacks.EarlyStopping(),
+                         tf.keras.callbacks.LearningRateScheduler()])
+
+```
+
+For complete list of blocks, please checkout the documentation [here](/block).
diff --git a/examples/functional_api.py b/examples/functional_api.py
@@ -0,0 +1,35 @@
+import autokeras as ak
+import numpy as np
+import tensorflow as tf
+from keras.datasets import mnist
+
+# Prepare the data.
+(x_train, y_classification), (x_test, y_test) = mnist.load_data()
+x_image = x_train.reshape(x_train.shape + (1,))
+x_test = x_test.reshape(x_test.shape + (1,))
+
+x_structured = np.random.rand(x_train.shape[0], 100)
+y_regression = np.random.rand(x_train.shape[0], 1)
+
+# Build model and train.
+inputs = ak.ImageInput(shape=(28, 28, 1))
+outputs1 = ak.ResNetBlock(version='next')(inputs)
+outputs2 = ak.XceptionBlock()(inputs)
+image_outputs = ak.Merge()((outputs1, outputs2))
+
+structured_inputs = ak.StructuredInput()
+structured_outputs = ak.DenseBlock()(structured_inputs)
+merged_outputs = ak.Merge()((image_outputs, structured_outputs))
+
+classification_outputs = ak.ClassificationHead()(merged_outputs)
+regression_outputs = ak.RegressionHead()(merged_outputs)
+automodel = ak.GraphAutoModel(inputs=inputs,
+                              outputs=[regression_outputs,
+                                       classification_outputs])
+
+automodel.fit((x_image, x_structured),
+              (y_regression, y_classification),
+              trials=100,
+              epochs=200,
+              callbacks=[tf.keras.callbacks.EarlyStopping(),
+                         tf.keras.callbacks.LearningRateScheduler()])
diff --git a/examples/io_api.py b/examples/io_api.py
@@ -0,0 +1,21 @@
+import numpy as np
+import autokeras as ak
+from keras.datasets import mnist
+
+# Prepare the data.
+(x_train, y_classification), (x_test, y_test) = mnist.load_data()
+x_image = x_train.reshape(x_train.shape + (1,))
+x_test = x_test.reshape(x_test.shape + (1,))
+
+x_structured = np.random.rand(x_train.shape[0], 100)
+y_regression = np.random.rand(x_train.shape[0], 1)
+
+# Build model and train.
+automodel = ak.AutoModel(
+   inputs=[ak.ImageInput(),
+           ak.StructuredInput()],
+   outputs=[ak.RegressionHead(metrics=['mae']),
+            ak.ClassificationHead(loss='categorical_crossentropy',
+                                  metrics=['accuracy'])])
+automodel.fit([x_image, x_structured],
+              [y_regression, y_classification])
diff --git a/examples/task_api.py b/examples/task_api.py
@@ -0,0 +1,12 @@
+import autokeras as ak
+from keras.datasets import mnist
+
+# Prepare the data.
+(x_train, y_train), (x_test, y_test) = mnist.load_data()
+x_train = x_train.reshape(x_train.shape + (1,))
+x_test = x_test.reshape(x_test.shape + (1,))
+
+# Search and train the classifier.
+clf = ak.ImageClassifier(max_trials=100)
+clf.fit(x_train, y_train)
+y = clf.predict(x_test, y_test)
diff --git a/setup.py b/setup.py
@@ -13,6 +13,7 @@
     keywords=['AutoML', 'keras'],
     install_requires=[
         'tensorflow>=2.0.0b1',
+        'keras-tuner',
         'scikit-learn',
         'numpy',
         'lightgbm',