minor typos

Author: mitchellspryn

AirSimDistributedRL/ExploreAlgorithm.ipynb

@@ -59,7 +59,7 @@
  "As you will soon see, the reward function defined here is very basic and doesn't take into account some important parameters. Can you point out some obvious considerations this reward function overlooks?\n",
  "\n",
  "> **Thought Exercise 1.2:** \n",
- "The next time you are out for a drive, take note of how things happening around you on the road (the behavoir of other vehicles and pedestrians, traffic laws, roadsigns etc), the state of your car (your current speed, steering angle, acceleration etc) and your mental state (the urgency of getting to your destination, your overall stress/frustration level etc) result in you making decisions on the road. Reinforcement learning is unique as it is inspired by the behavioral psychology of human beings and animals. If you were to wrtite a reward function for how you drive in real life, what would it look like?\n",
+ "The next time you are out for a drive, take note of how things happening around you on the road (the behavoir of other vehicles and pedestrians, traffic laws, roadsigns etc), the state of your car (your current speed, steering angle, acceleration etc) and your mental state (the urgency of getting to your destination, your overall stress/frustration level etc) result in you making decisions on the road. Reinforcement learning is unique as it is inspired by the behavioral psychology of human beings and animals. If you were to write a reward function for how you drive in real life, what would it look like?\n",
  "\n",
  "To compute our reward function, we begin by computing the distance to the center of the nearest road. We then pass that distance through an exponential weighting function to force this portion to the range [0, 1]. The full definition of the reward function can be seen below."
  ]
@@ -120,7 +120,6 @@
  "cell_type": "code",
  "execution_count": 5,
  "metadata": {
- "collapsed": false,
  "scrolled": false
  },
  "outputs": [
@@ -209,9 +208,7 @@
  {
  "cell_type": "code",
  "execution_count": 6,
- "metadata": {
- "collapsed": false
- },
+ "metadata": {},
  "outputs": [
  {
  "name": "stdout",
@@ -262,9 +259,7 @@
  {
  "cell_type": "code",
  "execution_count": 7,
- "metadata": {
- "collapsed": false
- },
+ "metadata": {},
  "outputs": [
  {
  "name": "stdout",
@@ -355,7 +350,7 @@
  "name": "python",
  "nbconvert_exporter": "python",
  "pygments_lexer": "ipython3",
- "version": "3.6.0"
+ "version": "3.5.2"
  }
  },
  "nbformat": 4,

AirSimDistributedRL/LaunchLocalTrainingJob.ipynb

@@ -93,7 +93,7 @@
  "cell_type": "markdown",
  "metadata": {},
  "source": [
- "Note that training the model from scratch can take up to 5 days on a powerful GPU. Using pre-trained weights, the model should begin to visibly converge after 3 hours of training. (Stats reported from a GTX 1080)"
+ "Note that training the model from scratch can take up to 5 days on a powerful GPU. Using pre-trained weights, the model should begin to visibly converge after 3 hours of training. (Stats reported from a NVIDIA Tesla M60)"
  ]
  }
  ],
@@ -113,7 +113,7 @@
  "name": "python",
  "nbconvert_exporter": "python",
  "pygments_lexer": "ipython3",
- "version": "3.6.0"
+ "version": "3.5.2"
  }
  },
  "nbformat": 4,

AirSimDistributedRL/LaunchTrainingJob.ipynb

@@ -126,9 +126,7 @@
  {
  "cell_type": "code",
  "execution_count": 7,
- "metadata": {
- "collapsed": false
- },
+ "metadata": {},
  "outputs": [
  {
  "name": "stdout",
@@ -176,7 +174,7 @@
  "\n",
  "In each of these folders, a subdirectory will be created with your experiment Id. \n",
  "\n",
- "If you use remote desktop to connect to the agent machines, you will be able to see the training code drive the vehicle around (be sure to give administrator privileges to run any powershell scripts when prompted).\n",
+ "If you use remote desktop to connect to the agent machines, you will be able to see the training code drive the vehicle around (be sure to give administrator privileges to run any powershell scripts when prompted). \n",
  "\n",
  "Training will continue indefinitely. Be sure to let the model train for at least 300,000 iterations. Once the model has trained, download the weights and move on to **[Step 3 - Run the Model](RunModel.ipynb)**."
  ]
@@ -198,7 +196,7 @@
  "name": "python",
  "nbconvert_exporter": "python",
  "pygments_lexer": "ipython3",
- "version": "3.6.0"
+ "version": "3.5.2"
  }
  },
  "nbformat": 4,

AirSimDistributedRL/RunModel.ipynb

@@ -43,7 +43,6 @@
  "cell_type": "code",
  "execution_count": 17,
  "metadata": {
- "collapsed": false,
  "scrolled": true
  },
  "outputs": [
@@ -102,21 +101,9 @@
  },
  {
  "cell_type": "code",
- "execution_count": 19,
- "metadata": {
- "collapsed": false
- },
- "outputs": [
- {
- "name": "stdout",
- "output_type": "stream",
- "text": [
- "Connecting to AirSim...\n",
- "Waiting for connection: \n",
- "Connected!\n"
- ]
- }
- ],
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
  "source": [
  "print('Connecting to AirSim...')\n",
  "car_client = CarClient()\n",
@@ -224,7 +211,7 @@
  "name": "python",
  "nbconvert_exporter": "python",
  "pygments_lexer": "ipython3",
- "version": "3.6.0"
+ "version": "3.5.2"
  }
  },
  "nbformat": 4,

AirSimDistributedRL/SetupCluster.ipynb

@@ -32,7 +32,7 @@
  "* **\"batch_job_user_name\"**: Choose a username\n",
  "* **\"batch_job_user_password\"**: Choose a password\n",
  "* **\"batch_pool_name\"**: Choose a name for your pool of machines\n",
- "* **\"batch_pool_size\"**: The total number of virtual machines you want to use in your pool (minimum 2). You will need one machine to act as the parameter server and rest will take on the role of agents. For example, if you want to distribute training across 5 VMs, you will use a batch pool size of 6. \n",
+ "* **\"batch_pool_size\"**: The total number of virtual machines you want to use in your pool (minimum 2). You will need one machine to act as the parameter server and rest will take on the role of agents. For example, if you want to distribute training across 5 agent VMs, you will use a batch pool size of 6. \n",
  "\n",
  "Before setting up the cluster, you need to set up an Azure File Share to host the executable and the script files. Let's begin by importing some prerequisite libraries. "
  ]
@@ -134,16 +134,15 @@
  "source": [
  "Now that we have all of the prerequisite files generated, the next step is to create the Azure File Share. We create the file share and upload all of the files under the */Share* directory of the downloaded data files. Inside this directory, there are two folders:\n",
  "\n",
- "* **AirsimOpenGL**: This folder contains the AirSim executable, as well as some helper scripts used to launch the exe.\n",
- "* **scripts_downpour**: This folder contains the actual scripts that will be executed during the batch job. For more information about these scripts, see [Step 1 - Explore the Algorithm](ExploreAlgorithm.ipynb)."
+ "* **data**: This folder contains data files used by the executable. You will explore the uses of these files in [Step 1 - Explore the Algorithm](ExploreAlgorithm.ipynb).\n",
+ "* **scripts_downpour**: This folder contains the actual scripts that will be executed during the batch job. For more information about these scripts, see [Step 1 - Explore the Algorithm](ExploreAlgorithm.ipynb).\n",
+ "* **tools**: This folder contains some auxillary tools used to set up the VMs (e.g. [AzCopy](https://docs.microsoft.com/en-us/azure/machine-learning/team-data-science-process/move-data-to-azure-blob-using-azcopy) and [7zip](http://www.7-zip.org/))"
  ]
  },
  {
  "cell_type": "code",
  "execution_count": 25,
- "metadata": {
- "collapsed": false
- },
+ "metadata": {},
  "outputs": [
  {
  "data": {
@@ -239,9 +238,7 @@
  {
  "cell_type": "code",
  "execution_count": 16,
- "metadata": {
- "collapsed": false
- },
+ "metadata": {},
  "outputs": [
  {
  "data": {
@@ -330,7 +327,7 @@
  "name": "python",
  "nbconvert_exporter": "python",
  "pygments_lexer": "ipython3",
- "version": "3.6.0"
+ "version": "3.5.2"
  }
  },
  "nbformat": 4,