Merge pull request #395 from alexbrillant/master

H1, H2, H3 Automl Hyperparameter Tuning

Author: alexbrillant

examples/Introduction to Automatic Hyperparameter Tuning.ipynb

@@ -15,7 +15,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# AutoML loop\n",
+    "## AutoML loop\n",
     "\n",
     "The step AutoML is a step that can execute any Automatic Machine Learning Algorithms."
    ]
@@ -24,7 +24,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## 1. Define your pipeline\n",
+    "### 1. Define your pipeline\n",
     "\n",
     "Consider multiple sklearn classifiers that each have their own hyperparms: "
    ]
@@ -102,7 +102,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## 2. Choose a validation splitter\n",
+    "### 2. Choose a validation splitter\n",
     "\n",
     "Let's pick the [ValidationSplitter](https://www.neuraxle.org/stable/api/neuraxle.metaopt.auto_ml.html?highlight=validationsplitter#neuraxle.metaopt.auto_ml.ValidationSplitter) which is enough for most machine learning problems:"
    ]
@@ -131,7 +131,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## 3. Define a the main scoring metric with [ScoringCallback](https://www.neuraxle.org/stable/api/neuraxle.metaopt.callbacks.html?highlight=scoringcallback#neuraxle.metaopt.callbacks.ScoringCallback)\n",
+    "### 3. Define a the main scoring metric with [ScoringCallback](https://www.neuraxle.org/stable/api/neuraxle.metaopt.callbacks.html?highlight=scoringcallback#neuraxle.metaopt.callbacks.ScoringCallback)\n",
     "\n",
     "Let's pick the accuracy score as our scoring metric: "
    ]
@@ -158,7 +158,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## 4. Add metric callbacks with [MetricCallback](https://www.neuraxle.org/stable/api/neuraxle.metaopt.callbacks.html?highlight=metriccallback#neuraxle.metaopt.callbacks.MetricCallback) (optional)\n",
+    "### 4. Add metric callbacks with [MetricCallback](https://www.neuraxle.org/stable/api/neuraxle.metaopt.callbacks.html?highlight=metriccallback#neuraxle.metaopt.callbacks.MetricCallback) (optional)\n",
     "\n",
     "Let's add sklearn f1_score as an additional metric callback: "
    ]
@@ -191,7 +191,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## 5. Select an hyperparams repository \n",
+    "### 5. Select an hyperparams repository \n",
     "\n",
     "For the sake of this example, let's pick the [InMemoryHyperparamsRepository](https://www.neuraxle.org/stable/api/neuraxle.metaopt.auto_ml.html?highlight=inmemoryhyperparamsrepository#neuraxle.metaopt.auto_ml.InMemoryHyperparamsRepository) because we don't need to persist trial files:"
    ]
@@ -237,7 +237,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## 6. Select an hyperparams optimizer \n",
+    "### 6. Select an hyperparams optimizer \n",
     "\n",
     "Let's pick the infamous Tree Parzen Estimator algorithm:"
    ]
@@ -274,7 +274,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## 7. Create, and launch AutoML loop\n",
+    "### 7. Create, and launch AutoML loop\n",
     "\n",
     "Create the AutoML step that will execute the training loop:"
    ]
@@ -312,7 +312,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Run 10 trials "
+    "## Run 10 trials "
    ]
   },
   {
@@ -2665,7 +2665,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "### Get best model, and predict"
+    "## Get best model, and predict"
    ]
   },
   {