Hi, everyone. In the previous video clip, we installed the Anaconda environment in which you can find Python files. Later I will show you how they are installed in detail. Now we need to install other packages, required for Python coding and AI coding, because one purpose of this MOOC course is that creating AI coding environment. You are working on an Anaconda environment. We need to learn a few Anaconda commands. Because if you remember of few and many Anaconda commands, you can easily install or uninstall packages. Surely, instead of the "conda" command, you can use "pip" another Python library installing command, you can use "pip". But because you are working in Anaconda environment, let me first introduce a few Anaconda commands that are useful for your coding. First, "conda list." Because we are using Anaconda here is prompt, ' > .' This is the prompt sign. Then you need to always put "conda." "conda" is a starting command for doing a certain task. "conda+blank space+list" means that "List all libraries installed in an environment". So, by using "conda list," you can check what kind of libraries are installed on your computer. The second command line is the "conda create -n ml_env Python=3.8." Python equals 3.8. Because we are using Python 3.8, we are specifying Python 3.8. If you do not give this version information, then naturally, the newest version of Python will be installed. But we are going to use 3.8 or 3.7. That's why we are specifying what version of Python will be used in coding. And also, here's "ml_env." It is an arbitrary name for coding. So, for example, think of why we are making this new environment. Because you may later, you may get involved in multiple projects, but because you are working with others in different projects, you may use different Python version. So, it is a way of creating an environment for each project. So, in this case, I gave this environment name. You can give another name that you want to use. So you don't have to use this one. But anyway, in this case, if you're following this course, probably using the same environment, name is better. So "ml_env" machine learning under bar env is another environment for a specific project. And "conda env list". Because as I said before, we can create many coding environments. So there could be many environment. If you want to know what environments exist, you use this command, "conda env list". Then all the environments, coding environments created on your computer will be listed. "conda env remove." This one, "conda env remove -n ml_env." If you want to delete "ml_env" by using this command, you can simply remove the exsting coding environment. And because you are working in a specific environment, you want to do coding work in the specific environment, then you need to get into that environment. We are using the "ml_env" environmnet. In order to get into that environment, you are using "conda activate ml_env". By using this command that you are getting into that environment. And you. do some coding work. But after finishing coding, work, you want to get out of that environment "conda deactivate". It is a way of getting out of an environment. "conda --version" if you want to know your Anaconda version, use this command "conda --version". If you want to know Python version in that case,also simply "python --version". By using this command lines,you can perform basic tasks of managing your coding environment. And if you want to get out of that environment, after command prompt, you use this "exit". I will show you how to do it later. So, this is a brief overview of what I'm going to do with a real computer. So, in order to install additional libraries additional libraries like "NumPy" or "TensorFlow," we need to use "Anaconda Prompt". So, we already installed the Anaconda 3. If you click this "Window button," then all appear as shown here. Then if you scroll down, you will find "Anaconda 3 64-bits". And under that list, you click on a contact prompter. you will click on "Anaconda Prompt". Then you will see this Anaconda Prompt, and it is "base" rigth now where you are standing. You are staying at base environment. It is "C:". "users" directory. On the user directory, currently it is "KAIST," but actually in my computer currently the username is "KLMS." So later if I open this Anaconda prompt, you will see "KLMS" username. So this is what you saw later. You will see later. And you are. Installing another Python with version 3.8 using "conda". You are installing separately Python for each environment, because each environment can be used for specific project. Surely, one environment can be used for multiple projects. But sometimes it may confuse you later. So, creating an environment for each project is an efficient way of managing your coding materials. So, in this case, we already installed Python when we installed Anaconda. That is installed in the base environment. Now we are creating another environment , "ml_env" with Python 3.8. So, this is the command. Then these are the installing process. I'm going to show you this later. Then now, after "ml_env" is created, you want to get into "ml_env" environment. In that case, as I said before, "conda activate ml_env." And then from the “base” environment, you are getting into "ml_env."It (base) will be changed into ml_env. And the environment also exists under your username. So now you are installing additional libraries. First, you are installing "TensorFlow".That's why "conda install TensorFlow". Surely, you can use "pip" instead of a "conda" command. In this case, we are installing the TensorFlow CPU version, not the GPU version. In most computers, personal computers, GPU is not installed. But as a beginner, you are going to use your notebook computer or a desktop computer without a GPU. So we are installing CPU vversion. You can do many things, actually basic neural network. You can run basic network with a CPU version notebook computer. No problem at all!. So, when you install CPU version, you simply type "TensorFlow". In case of GPU, you need to specify "--gpu". But we are not installing that. And in order to install GPU, you need to check your graphic card. So, it is a little bit more complex and difficult to install. But we are installing TensorFlow. And we are not giving specific version information of a TensorFlow. If you want to use TensorFlow 1.x.x something. In that case, you need to give a version number. But last year, TensorFlow 2.0.x was released. So you need to learn "TensorFlow 2.0". So, in this case, you don't need to specify version information of TensorFlow. Then, naturally the newest version of TensorFlow will be installed. So, after installing TensorFlow, you need to check whether TensorFlow is correctly installed or not. This is the checking way of TensorFlow installation. But rather than explaining it here, I will show you later. After installing Tensorflow, TensorFlow is for neural network, deep learning. Now, we need to install other packages. In the previous slide, we installed only one package. But surely, you can install multiple packages with one command line. "conda install," you simply elaborate the packages you are going to install "NumPy, Pandas, Matplotlib, Seaborn, Scikit-learn, Keras". You are installing other packages for machine learning and Python coding. There should be always "one space" between libraries. Then at one time, all packages will be installed. And if you want to check what packages are installed in your specific coding environment. Simply use "conda list". That's what I already explained. Now, what about you installed unnecessary packages? So, you want to remove a specific library or package. In that case, you simply use a "conda unistall" and package name without this one "<>". Simply, name package. Then that specific package will be deleted so you can remove and re-install a specific package. Sometimes, you choose an advanced or upgraded package. Then remove previous package. You can do it with those commands. And also packages often build on other packages. For example, "Pandas" was built on "NumPy" and "Keras" was built on "TensorFlow". So what? It means that some packages are depending on other packages. So in order to use "Pandas," you must have "NumPy." Without "NumPy," you cannot use "Pandas." Without "TensorFlow" you cannot use "keras" because two packages are depending on each other. So sometimes you need to install some packages together. But don't worry we already installed all the necessary packages. So you don't have to worry about it, but simply I am giving you information about the dependency among the packages. So, before showing you the full installing process, here's your review question: Which one is it correct for listing an environment that you created? There are five options. Multiple choices are provided. Which one is the answer? So, which one? Think about! Looks similar, but they are slightly different. There's only one answer. So, "conda env list." This one, right? So, the answer is three, "conda env list". If you want to check what are all the environments that you created, then simply use it. If you want to check them, you use "conda env list" in the command-line.