header formatting

Author: jeffjennings

2_user-defined-model.ipynb

@@ -224,7 +224,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Fit an Emission Line with a Gaussian Model"
+    "### Fit an Emission Line with a Gaussian Model"
    ]
   },
   {
@@ -289,7 +289,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Exercise \n",
+    "### Exercise \n",
     "\n",
     "Go back to the previous plot and try to make the fit work. Note: **Do not spend more than 10 minutes** in this exercise. A couple of ideas to try: \n",
     "* Is it not working because of the model we chose to fit? You can find more models to use [here](http://docs.astropy.org/en/stable/modeling/#module-astropy.modeling.functional_models).\n",
@@ -302,7 +302,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Compound models"
+    "## Compound models"
    ]
   },
   {
@@ -529,7 +529,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Exercise\n",
+    "### Exercise\n",
     "Modify the value of delta to change the minimum and maximum values for the mean of the gaussian. Look for:\n",
     "* The better delta so the mean is closer to the real value of the $H\\alpha$ line.\n",
     "* What is the minimum delta for which the fit is still good according to the plot?"
@@ -539,7 +539,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Custom model"
+    "## Custom model"
    ]
   },
   {
@@ -559,7 +559,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Basic custom model"
+    "### Basic custom model"
    ]
   },
   {
@@ -699,15 +699,15 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Exercise\n",
+    "### Exercise\n",
     "Modify the initial conditions of the fit and check yourself the relation between the best fit parameters and the initial conditions for the previous example. You can check it by looking at the Reduced Chi Square value: if it gets closer to 1 the fit is better and vice versa. To compare the quality of the fits you can take note of the Reduced Chi Square value you get for each initial condition."
    ]
   },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Full custom model"
+    "### Full custom model"
    ]
   },
   {
@@ -835,7 +835,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Exercise\n",
+    "### Exercise\n",
     "Play with the initial values for the last fit and improve the Reduced Chi Squared value. \n",
     "\n",
     "**Note:** A fancy way of doing this would be to code a function which iterates over different initial conditions, optimizing the Reduced Chi Squared value. No need to do it here, but feel free to try."
@@ -845,7 +845,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Exercise\n",
+    "### Exercise\n",
     "\n",
     "Custom models are also useful when we want to fit an **unusual function** to our data. As an example, create a full custom model to fit the following data."
    ]