Transcript:
Welcome back to this series on Nil Network programming with Pi Torch in this video, we’re going to cover the needed prerequisites for installing Pi Torch. It’s pretty easy to get up and running with high torch. So without further ado, let’s get it done. [MUSIC] Getting started with Pi Torch is relatively easy. The recommended best option is to use the Anaconda Python package manager with Anaconda. It’s easy to get and manage Python Jupiter notebook and other commonly use packages for scientific computing and data science, including Pi Torch. Of course, these are the steps to get PI torch installed. First we want to download and install Anaconda. Then we want to go to the Pi Torch website to the getting started section, so we can get the right command to run on our particular system once we have the command from the PI Torch website. We run the command at our terminal. I have all the links you need. On the text-based post on deep laser comm, we’ll first go to the Anaconda website here. You’ll just download and install the latest version, which is Python 3 6 at the current moment. Once you have Anaconda installed, we need to go to the Pi Torch website, and then we scroll down just a bit and we see the getting started section so here we select our configurations and then we’ll be presented with a command that we need to run at our terminal. So for my system. I’m running Windows. I’m using Conda As a package manager. I have Python 3 6 and I’m gonna go with q2 9.0 There’s no need to install cuda. Separately, they needed. CUDA software comes installed with PI torch. If we select a version of cuda here, I’m gonna go with q2 9.0 because I’ve seen some issues with 9.2 If you want to give 9.2 a shot, go ahead, try it out, See if it works. And if you run it in trouble, just uninstall it and then reinstall with the 9 dotto, so we’ll copy these commands and we’re ready to run them at our terminal. What you’ll notice is that we’re installing both PI torch and torch vision. I’m here now at my terminal inside Visual Studio code. So I’ll just paste the command and run the command. Conda, install Pi torch. – see hi. George, it’ll take a few minutes, so I’m gonna speed this up. We’ll go two to three X. So this is speedy now. Pi Torch is installed and we can just verify by running the command. Conda lists Pi Torch. We have a package named PI torch. The version is 0.41 and then we get a build and the channel looking at the bill closely. We can see the python version, the cuda version and the QD. And in version KU DNN is the deep neural network component of CUDA, hence the name Ku D and in so we’re ready now to begin working with Pi Tours before we do, let’s cover some of the software that we’ll be using in this series in this series, we’ll be using the following software for writing and debugging our code. The first piece of software is an integrated development environment called Visual Studio code. The second piece of software we’ll be using is an interactive environment called Jupiter Notebook. Once you have Visual Studio code installed, You’ll also want to install the Python plugin. This is done from inside vs. Code in the Plugin’s section. You can see. I have the Python extension. We’ll be using visual studio code. Primarily for debugging our code vs. Code makes debugging and inspecting our objects pretty easy. It’s also useful for exploring the Pi torch source code. The navigation features for navigating the source code are pre robust. We won’t use vs. Code until part 2 of the series, and most of our time will be spent inside Jupiter notebook. We automatically get Jupiter notebook with the Anaconda installation. Now, bear in mind, neither of these tools are necessary, but they do make our lives as developers a lot easier. I’m in a Jupiter notebook now before we run our first PI torch commands. Let me just show you the directory structure that we’ll be using for our project, So I’ve created a folder called Pi Torch, which you can see right here and navigating into this folder. We can see our project. Everything pi torch related is going to live here. You can see that we have a data directory. This is where our data will be located when we download that for our project for our convolutional neural network that we’re going to build. We have a resources folder where we’ll put some custom code files that we’ll write. And then you can see. We have two notebooks, part one part two and then the condensed version of the project, so I’m going to go ahead and open part one of the Neural Network programming series and we will run our first Pi Torch commands so here to verify the install. The first thing we’re going to do before we can run any. Pi Tortes commands is, of course, import the top-level PI towards package torch, so I will run this control enter and we have successfully imported torch so now we can use torch to ask for the version of Pi Torques that we have installed and as you can see the version we have is 0.4 dot one. We mentioned earlier that. Cuda comes installed with PI torch. So when we ask torch, CUDA dot is available to tell us whether or not CUDA is available on our system. We hope that the answer is going to be true and indeed, the answer is true. This is because we chose the 9.0 version of CUDA during the Pi torch install, and I have a supported NVIDIA GPU on my system. Finally, we can see or check the version of CUDA by calling torch version CUDA, so we’ll do that, and as we expect, we get 9.0 if your torch. Cuda dot is available. Call returns false. It may be because you don’t have a supported. NVIDIA GPU installed in your system So even if you chose 9.0 for the cuter version for the Pi Torch install, it’s still possible that you don’t have a supported NVIDIA GPU. And in this case, the call would return false, however, don’t worry, a GPU is not required to use Pi Torch or to follow this series. We can obtain quite good results in a reasonable amount of time, even without having a GPU. And when just starting out, I recommend not using the GPU. The goal of discussing a GPU here and in the next video is to recognize its availability so that we are prepared to leverage its capabilities when we begin tackling larger projects. If you’re interested in checking on whether your NVIDIA GPU supports CUDA, you can check for it on Nvidia’s website. The specific link is on the blog post on deep lizard calm. If you’re following the neural network series, be sure to watch the first two videos. If you haven’t already both of these come before this one in the series make sure to check out that deep lizard, hide mine for exclusive perks and rewards in the next video, we will learn more about CUDA GPUs, and importantly, why we even use GPUs in the first place. Let me know if you’re all set. Thanks for watching and contributing to collective intelligence. I’ll see you in the next One. Software engineers can only write so much software, but machines can write enormous software machine. It doesn’t get tired and it types very fast with GPUs. It can type very fast so long as there’s data so long as there’s knowledge in how to create the architecture creativity, we can create absolutely enormous software and this is the future of computing. [music] you [Music]!