diff --git a/README.rst b/README.rst
index 43281f7345f..e55caceac07 100644
--- a/README.rst
+++ b/README.rst
@@ -30,8 +30,8 @@ DALI is preinstalled in the `NVIDIA GPU Cloud <https://ngc.nvidia.com>`_ TensorF
 Installing prebuilt DALI packages
 ---------------------------------
 
-Prerequisities
-^^^^^^^^^^^^^^
+Prerequisites
+^^^^^^^^^^^^^
 
 .. |driver link| replace:: **NVIDIA Driver**
 .. _driver link: https://www.nvidia.com/drivers
@@ -62,8 +62,8 @@ Installation
 Compiling DALI from source
 --------------------------
 
-Prerequisities
-^^^^^^^^^^^^^^
+Prerequisites
+^^^^^^^^^^^^^
 
 .. |nvjpeg link| replace:: **nvJPEG library**
 .. _nvjpeg link: https://developer.nvidia.com/nvjpeg
diff --git a/docs/conf.py b/docs/conf.py
index 956de907145..810a3bcf982 100644
--- a/docs/conf.py
+++ b/docs/conf.py
@@ -41,11 +41,23 @@
     except:
         git_sha = u'0000000'
 
-# The short X.Y version
-version = str(version_short + u"-" + git_sha)
+
+
+git_sha = git_sha[:7] if len(git_sha) > 7 else git_sha
+
+version = str(version_long + u"-" + git_sha)
 # The full version, including alpha/beta/rc tags
 release = str(version_long)
 
+# hack: version is used for html creation, so put the version picker
+# link here as well:
+version = version + """<br/>
+Version select: <select onChange="window.location.href = this.value" onFocus="this.selectedIndex = -1">
+    <option value="https://docs.nvidia.com/deeplearning/sdk/dali-developer-guide/">Current release</option>
+    <option value="https://docs.nvidia.com/deeplearning/sdk/dali-master-branch-user-guide/docs/">master (unstable)</option>
+    <option value="https://docs.nvidia.com/deeplearning/sdk/dali-archived/index.html">Older releases</option>
+</select>"""
+
 # -- General configuration ---------------------------------------------------
 
 # If your documentation needs a minimal Sphinx version, state it here.
@@ -59,6 +71,8 @@
     'sphinx.ext.autodoc',
     'sphinx.ext.mathjax',
     'sphinx.ext.napoleon',
+    'sphinx.ext.ifconfig',
+    'sphinx.ext.extlinks',
     'nbsphinx'
 ]
 
@@ -183,3 +197,6 @@
 
 
 # -- Extension configuration -------------------------------------------------
+extlinks = {'issue': ('https://github.com/NVIDIA/DALI/issues/%s',
+                      'issue '),
+            'fileref': ('https://github.com/NVIDIA/DALI/tree/' + (git_sha if git_sha != u'0000000' else "master") + '/%s', ''),}
diff --git a/docs/examples/getting started.ipynb b/docs/examples/getting started.ipynb
index b9b4fb241fb..264844c602d 100644
--- a/docs/examples/getting started.ipynb	
+++ b/docs/examples/getting started.ipynb	
@@ -20,6 +20,15 @@
     "It is a single library, that can be easily integrated into different deep learning training and inference applications."
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Optimal configuration\n",
+    "\n",
+    "DALI offers ease-of-use and flexibility across GPU enabled systems with direct framework plugins, multiple input data formats, and configurable graphs. DALI can help achieve overall speedup on deep learning workflows that are bottlenecked on I/O pipelines due to the limitations of CPU cycles. Typically, systems with high GPU to CPU ratio (such as Amazon EC2 P3.16xlarge, NVIDIA DGX1-V or upcoming NVIDIA DGX-2) are constrained on the host CPU, thereby under-utilizing the available GPU compute capabilities. DALI significantly accelerates input processing on such dense GPU configurations to achieve the overall throughput."
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
diff --git a/docs/examples/pytorch/pytorch-resnet50.rst b/docs/examples/pytorch/pytorch-resnet50.rst
index 901962997d5..47a56918e70 100644
--- a/docs/examples/pytorch/pytorch-resnet50.rst
+++ b/docs/examples/pytorch/pytorch-resnet50.rst
@@ -30,7 +30,7 @@ Requirements
 Training
 --------
 
-To train a model, run `main.py <https://github.com/NVIDIA/DALI/blob/master/docs/examples/pytorch/main.py>`_ with the desired model architecture and the path to the ImageNet dataset:
+To train a model, run :fileref:`docs/examples/pytorch/main.py` with the desired model architecture and the path to the ImageNet dataset:
 
 .. code-block:: bash
 
diff --git a/docs/index.rst b/docs/index.rst
index c810f188966..df093208658 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -10,8 +10,15 @@ NVIDIA DALI documentation
    :start-after: overview-begin-marker-do-not-remove
    :end-before: overview-end-marker-do-not-remove
 
-.. warning::
-   You are currently viewing unstable developer preview of the documentation. To see the documentation for the latest stable release click `here <https://docs.nvidia.com/deeplearning/sdk/index.html#data-loading>`_
+.. ifconfig:: "dev" in release
+
+   .. warning::
+      You are currently viewing unstable developer preview of the documentation. To see the documentation for the latest stable release click `here <https://docs.nvidia.com/deeplearning/sdk/index.html#data-loading>`_
+
+.. toctree::
+   :hidden:
+
+   Documentation home <self>
 
 .. toctree::
    :maxdepth: 2