TensorFlow for Deep Learning course Udemy launch Q&A (live)

Here we go. It looks like we're ready. Who's in? We're going live. This is so exciting. I need my pop-top. All right. So, you know the drill with this. The first things first, we start every live stream with a little dance because it is so exciting that we are launching the TensorFlow for Deep Learning course on Udemy. This has been a long, long time coming, and that's the purpose of this stream, is to answer any and every question that you have, particularly about the course, That will be the focus of the stream, but of course, feel free to ask me anything. There are a lot of new people to the channel, so welcome, welcome to the course. Welcome to the channel. Welcome, welcome, welcome wherever you are in the world. I am streaming to you live from Brisbane, Australia. So Southern Hemisphere, you may live in the Southern Hemisphere, you may live in the Northern Hemisphere. It's amazing to have you here. Let's have some fun. And now of course we always need, if you let me know, how is the sound? How are the visuals? If everything is working, give me a thumbs up in the chat and start the questions rolling in. I've got a little screen set up here and I've got the chat over here and I've got the questions. We're going to start rolling with them as soon as they start rolling in. So whatever you want to ask, put it in here. But let's celebrate. You ready? I'm gonna do this away from the microphone, but to the TensorFlow for Deep Learning course on Udemy! Woo! Who's ready? Who's ready? Who's ready? Oh, I'm so excited. So, so excited. Now, my live chat, All right. Welcome. We've got about 30 or so viewers. Who wants to see a dance? We'll do the— who remembers Ziz? Remember this one? Where we're like— like where you're at a music festival. You just go— just cut shapes, you know? And my friend once told me, this is a real good one, it's all about the wrist movement. So if you do your wrist, right, you want to move your wrist, move your hips, it's like all joints are flowing into one. From the ankles to the hips to the wrists, just get around. Oh, thank you for the thumbs up. Okay, it looks like we're about 30 seconds or so behind, which is perfect. Perfectly fine. Jackson, how are you? Michael, how are you? Ashik, how are you? Addie, how are you? Marsha, Woodlegs, Krish, Gaurav, PC Nerd, Saket, Gaurav, everyone. Who has some questions about the course? Or do you just want to see me dancing for, for an hour? Because I can do that. I need to do some movement. I haven't, I didn't go for a walk this morning, contrary to what I usually do. I haven't gone for a walk. Oh, we have a question from Zafar. Can this course even help if I don't know anything about TensorFlow? Here we go. We've got some questions rolling in. Zafar asked, can this course help even if I don't know anything about TensorFlow? Yes, this course is designed for beginners who know nothing about deep learning or TensorFlow and will teach you both. So it's a beginner-friendly course for TensorFlow. If you've had some— well, let's just have a pre— have a look at the prerequisites on GitHub. I'm going to switch to my screen. So there's the first question. Thank you from Safar. Safar. And so here's the prerequisites. So what do I need to know to go through this course? If you have 6 months of writing Python code, right? Can you write a Python function? Have you done at least one beginner machine learning course? So my Beginner Friendly Machine Learning course is a great prerequisite to this course. That's a plug for me. Are you comfortable using Google Colab or Jupyter Notebook? So if you can use either of those, you'll be able to do this course. If you can, if you've got these 3, you'll be able to do the course. Though I would encourage you to try it anyway because I've kind of set it up in a way to be as approachable as possible. But should you do this course? Do you have 1+ years experience with deep learning and writing TensorFlow? If yes, no, you shouldn't. Use your skills to build something. This is— I can't stress enough, this is a beginner level course. If you've got some experience over a year in the deep learning field, this is probably not for you. So that's, that's a big, big, big, big upfront. But let's go back to the chat. What do we got here? Is this how you start your day with coding? Nadiru asks. Yes, I totally do. Nadiru, I usually start my day with some sort of physical exercise. I love to get outside. You can't see it here, but out there. I can see some beautiful trees, I can see the water. I love to go for a walk because it gets my mind into the game. So I start with physical activity, some sort of writing, whether that be code or words. So that's how I like to start my days. And of course, that's always changing. Um, Krish says, how to get started in machine learning? Here we go. Thank you for the question, Krish. Uh, how to get started in machine learning from zero to advanced? Well, My answer here is I could always just say, do my beginner ML course. But the truth is, just follow your curiosity. Right? So the internet, you've got the internet, Krish. If you could search online, which you can, you found this video, you can get started with machine learning. There is no set way of how to do something properly or how to not. I've got a video that's a roadmap. You could follow that. You don't have to follow that. It's just whatever you want to get into. Realize that you can follow your curiosity. However, to gain skills in anything, including machine learning, that requires discipline. So you need to practice what you learn. So that's the main two things. Follow your curiosity. And then once you've found something, have the discipline to keep following it. But the truth is, I've got a great tweet on this, by the way. Let me show you my screen. So this is Chris's question, how to get started in machine learning from zero to advanced. So you could go Daniel Bourke machine learning roadmap. I've got a great article on that. So here's one option. Uh, and then the other one, which I would actually recommend, screw all the roadmaps themselves, go to my Twitter. It's one of my recent tweets. I need to just log in here and oh, my screen is busted. Actually, I'm not even going to get distracted. I'll get distracted on Twitter. But yeah, one of my recent tweets, which is, uh, just basically, uh, use the internet. Search what you're after and then go down the rabbit hole for like 6 months and figure things out. Okay, here we go. Uh, Meet Gandhi, is this course— has a question here. I'm just going to keep going through these questions as many as I can for the next half an hour, hour or so. So let me know if you're enjoying them in the chat. Um, just keep asking a question. If I don't see it, just post it again. Is this course version TensorFlow version 1 or 2? It is TensorFlow 2. It's actually 2.2 plus, but we use features from TensorFlow, TensorFlow 2.4 plus, which is mixed precision. So right now the latest version of TensorFlow is 2.4. That's what the course will cover, 2.4 plus. But it actually starts out at 2.2 plus. So it's definitely version 2 onwards. And then if any amazing new features come out in TensorFlow, I'll be sure to let you know through the GitHub or through future tutorials or something like that. Oh, Gaurav, you got COVID, my friend. Damn it. Well, thank you for sharing that with me, and I wish you all the best and good health, my friend. Zhang Feng asks how to— will I teach how to build an ML pipeline on AWS? No, my experience is mostly with GCP, so I'm going to stick with that. However, AWS steps are very similar. So I have a video on my YouTube channel, you can look that up, which is how to, how to deploy a machine learning model using Google Cloud. That's just because I have the most experience with Google Cloud. AWS and Google Cloud and Microsoft Azure are very similar. They just have different names for their product service offerings. So for example, Google Storage is actually S3 in Amazon, which is I think is Simple Storage Solution, S3. But AWS has some great guides on how to build ML pipelines. If you want to learn things about the cloud, I would look up acloudguru.com. Here's how I learned about a bunch of stuff from the cloud. So right here, I'll go— if you want to learn about the cloud See here, A Cloud Guru, great website on there. Otherwise, a lot of practice. That's with all these things, right? There's a lot of practice. Here we go. What are we up next? Oh my goodness. Thank you so much everyone for joining in, by the way. I really, really appreciate this. As I said, I'm going to answer as many questions as I can, and then I can share this as well. So what's the date today? It's the 4th. It's the 30th of April. How crazy is that? We're already a third, a third of the way through 2021. Okay, so Sai asks, Sai asks, will it completely teach us to qualify for the TensorFlow Developer Certification? So yes, that's what, uh, the course materials are based on. I got certified during 2020, then started building this course. Um, many of the TensorFlow Developer Certification materials are in the course, and extra resources will be linked for practicing what else is needed. Okay, so I can't stress enough that although the course is called TensorFlow Developer Certification, I'm calling it TensorFlow for Deep Learning. I mean, that's what the GitHub repo is called. SWB Audio. In essence, this course teaches you 3 things. It teaches you the concepts of deep learning, It teaches you how to write TensorFlow code to, to learn those concepts. And both of those two together, those first two together, will prepare you for the TensorFlow Developer Certification. I'm just going to correct this spelling here. Right. Now, the certification is completely optional. You don't have to do that. In fact, I would argue that the skills themselves of TensorFlow, of how to— of deep learning and TensorFlow, just being able to apply them is more important than the certification. You can create your own version of a certification by creating nice projects on GitHub and sharing them with the world, having a good blog, having just a resource online where people can see what you've done. And that's what you can do when you know deep learning concepts as well as TensorFlow code concepts. The certification is a nice to have, not a need to have. I'm going to write that down here. The certification— note. Is a nice to have, not a need to have. So yes, it will help you, but it is not 100% necessary. Shashank says some of the course contents aren't uploaded yet. Yes, he's 100% correct. Are not uploaded yet in Udemy. So yes, You're right. There are, there's a few things missing, right? A few things missing. A few videos missing, but they are coming. So right now there's about 70 videos, 70 extra videos. So the course has been released on Udemy, right? With about, let's, let's just have a look. Hey, deburke.link. So this is where you can sign up. If you want to. There's a really great deal going at the moment, probably the best deal that the course is ever going to be. So deburke.link/udemytflaunch. Let's have a look at this. Um, so look at that deal. That's Australian dollars. That's $11 in my currency. That's a phenomenal deal. Frequently Bought Together. Andre. Oh, look at these two amazing instructors. Alright, where is the materials? Here we go. So this is what's coming soon. It is, oh, 10 more sections. Look how big this course is. Okay, so right now NLP Fundamentals, more videos are coming soon. I've recorded videos for these 2 sections, so they'll be coming, right? These 2 as well as, uh, these 2 sections, they'll be coming next. So these 2 sections are done. We're just editing them and uploading them now. Time Series Fundamentals in TensorFlow and Milestone Project 3. They'll be done by the end of May. And same with this. So these 3 sections will be done by the end of May. And these are done already. They'll be done any, any day now because they're just— we're just editing the videos, waiting to upload them. However, there is still 40 hours worth of videos already there ready to go. So that's more than enough to get started. So Yeah, if you want to sign up, go get the best deal you're probably ever going to get on Udemy and all the material, everything that's listed here is included. So you're not going to pay any extra when you sign up. You get everything including the future stuff that we upload. So just keep that in mind. You sign up now, you have the course for life. If you don't like it, well, you've got 30 days or something to do a refund. That's perfectly fine. All right, let's come back. Uh, right now it's about 70 videos. Waiting to be edited. They will be up as soon as possible. Everything will be finished before the end of May. All right, where are we on? Uh, Zafar says, are there any coupons for this course today? Yes, the best course, the best deal you're going to get is— I'm going to post the link in the chat right now. It's wb-audiocoaching.com.au. TF Launch. That already has the best code applied, so go do that. Michael says, side question, been meaning to get an electric stand-up desk, adjustable desk. Any recommendations from yours? So Michael, I think they're all mostly just the same these days. They're all pretty good. So mine goes up and down. I mean, that's what you want, right? So I think mine is just a generic one that you can buy from many office furniture stores. For me, I've decided when I want to buy something, I'm just going to buy it because otherwise you just spend way too long comparing different things, especially for something like a standing desk. If it goes up and down, that's, that's what I'm after, right? All right, where are we at? Okay, the chat just skipped me ahead. So Gaurav says everyone liked the video. Let's get it to 100 likes. Great idea, Gaurav. Get this video to 100 likes, people. And you don't have to, of course, only like it if you're having fun. Uh, I recently, so here we go. So Sai Chandara, Am I saying that right? No. Sai Chandra says, I recently got certified as a TensorFlow developer and I felt it is very easy. So was it worth it? If you know TensorFlow, you'll probably find the certification easy. So the point still stands. The certification is not 100% necessary. What is more necessary is building skills and making things with those skills. SWB Audio Capture now registered. And then sharing your work. GitHub/blog. I will say this till the cows come home, right? Certificates are nice to have. I'm going to move my cup out of here. Certificates are nice to have, right? Nice, nice. Not need to have. What companies are looking for are people with skills. So I believe comma.ai has a great comma.ai. This is one of my favorite things. If you go to comma.ai jobs, what we are looking for— here we go— contributors. So people who have contributed to open source projects. Show us your GitHub. Got stars? Have you built things? This is literally, they're asking you, what have you made? That's GitHub, right? Certifications are one form of showing your skills. Making things on GitHub is another form of showing your skills. And I would argue that for me personally, if I was looking for someone, the first thing I'm going to do if I'm trying to hire someone, I'm not looking to hire anyone right now, so just, this is just theoretical. Not in practice, would be I would go to their website, I would go to their GitHub, I would see the projects that they've worked on. If they have a certification, that's nice, that's really cool. But if they've worked on this project, wow, that's even better, right? So certification, nice to have. Website, things you've worked on, need to have. So this is who they're looking for. People who have contributed value to open source projects. Show us your GitHub. Have you got stars? Code is never measured in quantity, but rather quality. Of course, in the beginning, when you're first learning to code, write as much as possible. But as you get better, you will start to write less code because you're thinking more about the code that you're writing. Right? You're not just going out there and writing heaps of crappy code. You're thinking, okay, how can I best do this so I only have to do it once and rewrite it a little bit rather than having to rewrite it lots and lots. But when you're first starting out, just to get used to it, write as much as possible. So, SWB Audio Capture, that's not registered. This is my recommendation. So, uh, Sai Chandra says, I recently got— so this is what I'm saying, if you know TensorFlow, you'll probably find this certification easy. So if you're already good at TensorFlow, certification will probably be easy. If you know zero TensorFlow you'll probably find it hard. Does that make sense? But use your TensorFlow skills and your deep learning skills to make things and share your work. Here we go. Where we at next? Paul Garcia. Have I done anything, any cool things with ML recently? I'm going to put this in here. And no, I haven't, other than creating a course. So no, I haven't. My complete focus has been on the course, trying to get it finished. After the course is 100% finished, I'll be working on my own cool things with ML. So courses take a lot of work. Right? You have to do things like 4 or 5 times because first of all you have to learn them, then you have to use them, then you have to make materials for them, then you have to teach the materials for them, then you have to troubleshoot the materials. So that's like 5 times. So all of my efforts have been dedicated to that because otherwise I wouldn't be— if I was doing other things while I have students like yourselves, people who are watching this stream who are enrolled in the course and I'm doing other things instead of helping them fulfill the thing that I've promised them, people who have paid me to enroll, I would be basically breaking their trust. So my 100% focus right now is finishing this course, nothing else. I've got 2 projects that I'm going to work on after it's finished. I'm very, very excited for those. I'm not going to say anything else, but stay tuned to the channel. You will see them. Okay. Do we need any experience in ML? So from Adi, Adi Kumar, do we need any experience in ML? So if you do, Adi, that would be great. Yes, the formal prerequisites, prerequisites. How do you spell prerequisites? Prerequisites. Say at least one beginner ML course is required, but try the TensorFlow videos on my channel and see how you go. So I've got the first 14 hours of the course on my channel, so if you, if you want to try that out, you can go to part 1, right? So there's 10 hours of TensorFlow videos I'm going to go here. See video 1, 10 hours of TensorFlow. If you can go through this video, you can definitely do the full course. Otherwise, if you want to see the formal prerequisites, which I've put on the GitHub page prerequisites, here we go. I'm going to copy that link. What do I need to go through this course? I'll put it here. I've got see full prerequisites. Boom. So yeah, but if you can go through this, you can definitely go through the whole course. So I'd give that a try. But if you've done— you'll need some Python knowledge because I just start coding Python straight away. And if you know some ML terminology, like the concept of a training, validation, and test set, very important in machine learning. We do kind of touch on that in the course, but a lot of it is kind of assumed knowledge. So just give it a go and see how you find it. Here we are. SWB audio capture not registered. SWB audio app is running. Hmm, is the audio okay? Give me a thumbs up if you can still hear the audio. Everything sounding okay? I've got a little message from Ashik in here saying that their audio might be a bit busted. Uh, Nadiru asks, I am learning web development, uh, now, and later on I'll start ML or AI because I think that needs to have some experience on programming. Am I right? Um, so there's two different Of course. Yeah. So ML or AI and web development are kind of two different paradigms of programming. Now, one you would, you could consider software 1.0, the other you could consider software 2.0. So in web development, you're building a website and you're giving a bunch of rules and you're writing the website, do this. Whereas machine learning, you're writing the inputs and outputs and you get the machine learning algorithm to learn the difference between the input and outputs. So you don't necessarily have to learn web development for AI and ML. However, if you have some experience with programming, especially Python programming, that will probably be very helpful when you learn machine learning. So web development is not 100% required for AI/ML. However, AI/ML involves plenty of programming, so experience there would be valuable. Um, especially not registered— now it depends, right? It depends what kind of programming you're doing, but the, the language of machine learning and AI at the moment is Python. Python programming. Now, of course, you could use C or C++ or whatever, but if you're a beginner, Python is probably the most friendliest for you to get started. Oh wow, Ashik, thank you so much. You know what? Let me fix this, alright? System Preferences. I got a live— apparently I installed this app so it could record my audio, and I need to can it. How about I do this? Mute. And I'm gonna mute it as well. Okay, so I think that may have fixed it because I downloaded this app to help me record my computer's audio as well as my microphone's audio. So let me know if that's gotten rid of it. I think, I think it, it should have, but I, and the fact that it's telling me it's not registered, they can also get stuffed because I paid for their app. I paid $12 for their app. And it worked, but now it's telling me it's not registered. So thank you. And will the course cover using TensorFlow on Google Cloud? Let me put in this. So great question from Judit. No. So no, this is a beginner style course and focuses strictly on getting familiar with foundational TF concepts. If you want to have some practice using TensorFlow with Google Cloud, you can watch my video on how to deploy a machine learning model on Google Cloud. Otherwise, I would recommend— so if we go here, Coursera, Google Cloud— I'd recommend all of Google Cloud courses on Coursera. So heaps of different options there. So if you want practice on Google Cloud, see Coursera's offerings. Um, I'm gonna make a page in a second, or probably in a week or so, maybe throughout, just exactly what the course covers. So the GitHub is pretty vague in terms of what exactly gets covered. There's a lot of notebooks here. So this is the course materials built in there. But I'm going to— like the exercises and extracurriculum section, I'm going to make a page which just says what exactly each notebook covers. That way we'll have a quick overview of— I can send you here, hey, this is what the course covers. So it doesn't cover anything on Google Cloud aside from how to use Google Colab. I have a video on how to use Google Cloud to deploy machine learning model. That video is actually based off models we build in the course. So that would be a great extension. Okay. Daniel Bourke, Supernova. So So Supernova says, um, will you ever do any classes with programming in R? There's a lot of ML jobs in my area that require R. Um, not planned for now. My experience is with Python, not R. So no, I don't have any plans for R in the future. I've never used R. Not that I've got anything against R, I'm just— my experience with Python. So I can only teach the things that I have experience with. Yogesh, how are you, my friend? Good to see you. Course is going great with the— sorry, you're going great with the course. It's great to hear. Thank you so much for tuning in. Glad to hear you're enjoying it. Um, Rashu Tagi. So is TensorFlow 2.0 and PyTorch similar to each other? Mostly. So yes, they are more so. Previously, before TensorFlow 2.0, I would say that PyTorch and TensorFlow kind of had different paradigms. But now that TensorFlow has gone to 2.0, they are more similar than they are different. For example, if you're experienced in TF, you can get up and running in PyTorch in a few days, weeks, and same as vice versa. So the main paradigms between them both, right, that's the deep learning frameworks that kind of now got the best of both worlds from each other. Is that in my experience using both of them, I have more experience with TensorFlow, hence why the course is on TensorFlow. But the paradigms, if you know the overall concepts of what you're trying to build, you're trying to— what are you trying to do with machine learning? You're trying to get data from one place to another and you're trying to write some sort of manipulation algorithm in between. So you need to load the data, you need to find patterns in the data, and you need to export your model at some form. So in terms of doing both of those, yes, they are now both very similar. In PyTorch, you've had like custom training loops and evaluation loops for a long time. TensorFlow sort of has more of a— how would you say— a straightforward version with the fit method, which we use a lot throughout the course. But if you want to now write your own custom training loops in TensorFlow, you can do the exact same thing. So that's just something to keep in mind. They are, they do both. They're both amazing. And now I would say within the last 12 months they've, they've started to converge in terms of how they do things. All right. Yogesh asks, please, after this course, do you plan to make, um, TensorFlow deployment course. So no plans for a next course. I'm going to be working on my own projects for a while, which will probably include deployment. In the meantime, check out my video on Deploying models with Google Cloud. So I have a video, of course. Daniel Bourke, um, machine learning deployment. Here we go. This is my video. Go check this out if you want to learn about Deployment. That should get you started. But I have no plans for another course after this one. My 100% focus right now is finishing this one. It is not finished yet. As I said, we've released most of the videos. There's still about another 50 to 60 to 70 or so videos to come. Yeah, here we go. So there, right now, there's 70 videos waiting to be edited. They will be up as soon as possible. Invalid says, oh, this is a great question. What can you do after the course? So I have a little section. I'm going to probably expand on this a little bit more in a minute, but I would say number one, follow your curiosity, find something to work on. Which excites you and share your work. Two, uh, for a more formal list of exciting places to go check out. So there's a section on the GitHub now which has extensions. Of course, there's a lot of challenges throughout the course to go for, but possible places to go after the course. So this is extensions. This is on the GitHub README. Check this out. I recommend the deeplearning.ai specializations. If you want a reference book to read, I recommend Hands-On Machine Learning with scikit-learn, Keras, and TensorFlow. I've got this, one of my favorite— well, my favorite book on TensorFlow right now. So I would say this book— my course is a great foundation setter. This book is great to have as a reference and read after. If you want to learn how to build full-stack deep learning applications, otherwise, in other words, turn the models that we use in the course into things that people can use, they can get their hands on, check out Full Stack Deep Learning. I'd also recommend fast.ai. And if you want to get experience on how to get a job in the field, go and read my article on how can a beginner data scientist like me gain experience. So this is my best advice for getting any kind of job. But caveat, I am not looking for a job. So always take the advice from someone who is not actually doing what they say they're doing in terms of when I give you advice for you to look for a job, you must know that I am not looking for a job. So always take that. Uh, I'd rather build my own. So if you want a job, follow this. This is my best advice there. Michael asks a great question. What projects will be worked on throughout the course? So this is a little spoiler alert. There's 3 milestone projects throughout the course. One will be Food Vision, will be build an image classification model to recognize food. Also beat the DeepFood paper. Number 2 will be SkimLit. So replicate an NLP machine learning paper to organize medical abstracts. And number 3 will be BitPredict, which is coming soon, but it is use deep learning to predict the price of Bitcoin. Bitcoin is not a capital C, is it? So this is time series, this is NLP. And this is Computer Vision. So they're the 3 projects. You can see the first 2 in the GitHub right now. If you go to the GitHub, you'll see Milestone Project 1, Food Vision, number 7, and you'll also see Milestone Project 2, Skim Lit. This is coming soon. I'll upload it once it's ready. To the GitHub and the videos for 07 are out. Uh, the videos for 09 have been recorded and are being edited right now. So they will be up probably within a week or so. Andrew Yang says, just want to say thank you for making up this course. Thank you so much, Andrew. I really appreciate it. So, uh, So great to see you here. Samantha says, any tips for becoming a freelancer in AI? Simply, Samantha, read this article. So, oh, I'm going to actually copy this question. So simply, Samantha, any tips for becoming Freelancer in AI is my best advice for getting experience. Experience/work in the field is contained here. So rather than me say everything again, I'll let you, you read through that article. By the way, I'm going to be sharing this Notion document. And once, once the stream is finished, so this will be in the description. But go read that. It's my article, How Can a Beginner— I'll share this here. Uh, Eddie says, do we need GPU to learn? Do we need GPU to learn? No, if you want to run the code, uh, locally. Yes, but if you're following through with the course, we use Google Colab, which offers a free GPU. Thank you, Google. All right. So throughout the course, everything is based— so if you click on any notebook in the course, such as TensorFlow Fundamentals, every notebook has a little button inside. And you'll see this throughout the course, right? This is, this is what we do. You'll get very familiar with doing this in the course as we go through the videos. You can click open in Google Colab and this will open up in a tool called Google Colab, which is a Jupyter notebook but hosted by Google. Okay. Um, all right. So there's probably some sort of login issue here with my Google Colab, but trust me, that will work for you because, oh yeah, I deleted all of my history. Before this stream,.google.com, but we use Google Colab. So if I log in here, colab.research.google.com, and I click New Notebook— again, we cover a lot of this in the course, you get very familiar with this— but this is how we access a free GPU. And we can go Runtime, Change Runtime Type, Hardware Accelerator. We can select the GPU, and this Google Colab instance So this, this Google Colab, we're using it through a website right now, but that is being hosted on Google's cloud. And they're giving us a computer to use that uses a GPU when we select that. So that's just something to keep in mind. I'm going to leave that now, come back to the livestream. Here we go. What are we at? E-Blade45 says, does the TensorFlow certification exam have financial aid? I don't know. I don't— I'm going to tell you straight up when I don't know. I don't know. But remember, the TensorFlow exam is optional. And, and not a requirement. Making your own things on GitHub, slash, your own blog are free and just as good, slash, even better than a certification. So just keep that in mind. GitHub is free. You can use GitHub for free. We've got 100 viewers. So big shout out wherever you are in the world. Hey, let's do a fun exercise. I just thought of this. Post your— wherever, whatever country you're listening from or you're tuning in, post your flag emoji in the chat. I want to see what countries we've got. And if, if I don't know, well, post your flag emoji plus the name of your country because I don't know all the flags. But yeah, eBlade45, if you can't afford the TensorFlow certification exam, remember it is optional. GitHub and your own blog are free to create on. So use free tools where you can. Yo-yo, here we go. Yo-yo asks, how much average time to learn deep learning for a beginner to advanced self-taught? So it really depends, right? How do you scale knowledge? How do you know when someone's a beginner, when someone's an advanced? I've been learning machine learning and deep learning for almost 4 years now, and I would still consider myself a beginner in some things, but I certainly know a lot more than when I started. So very hard to measure, but you can get very good from not knowing anything in 1 year. With dedication. So what I mean by that is, so from knowing zero, your first year you're going to have a lot of gains, but then remember after the initial gains, subsequent gains, uh, come slower. So that's the thing with skills, right? You might see a very big initial burst with any new skill, But then after that, it takes a while to refine those skills and really master your craft. So, as I said, I'm 3.5 to 4 years into this and I would still consider myself— I know I know a lot more than when I started, but I also know how much I don't know. I would never ever call myself an expert ever. All I know is that it makes me very curious and it makes me excited to learn. And I'm following that excitement and curiosity. Here we go. We got some flags. Look at this. Peter Cox is from Orange, New South Wales. Hello from Australia. We have Simply Samantha is from India. We have Josh Cummins from Southern California. Hello team. Um, we have Baba Malik from India. We have Coding Mastery from India. We have Brandon Hating from, uh, South Africa. I believe that flag is India, Canada, Australia. Yes. Australia, Indonesia. Uh, Stephen Walk. Hello. Says, I would like to say hello. Hello, Stephen. Sean Richards is from Vancouver, Canada. Nice to see you. Supernova is from Colorado. Wow, this is so amazing. We have people from all over the world. Look at this. We have multiple different time zones, team. This is so exciting. And 105 people joining in. Welcome. If you're new to the channel, welcome. I teach you machine learning on this channel. We learn machine learning and we learn how to get healthier and we learn, or we try to learn about how to think about the world. So if you want to stay for that, you're more than welcome to join in. Yash says hello. We have someone from Algeria. Here we go. Pratham says how to prepare for machine learning internship. So Pratham, I'm going to give you my, my article up here, which, which says how to get experience. So there's no way to 100% prepare. Have an open mind and be willing to learn. Never be afraid to share what you don't know. The smartest people are always willing— well, not always, not always, are most of the time happy to share their knowledge with you, right? Otherwise, read this post. It contains some of my best advice experience. Boom. Hello from Palestine. Wow. Indonesian. Oh, Selamat Pagi. I used to learn Indonesian in high school. How do I say Satu Tua Diga Empat Lima Enam Sepulu? I'm lost. Dalaan. Okay, Invalid asks another great question. Mexico, West Orange, New Jersey, Fiji. Wow, this is so amazing, guys. This is really warming my heart. Look at this, right? If we got the maps, we get the Maps app out. Look at this. Coding Mastery says, "Is M1 Mac mini 16GB RAM enough for deep learning?" Yes, it is more than enough to start coding. For running large deep learning models, you probably want a GPU. So maps.google.com. Look at us, guys. How exciting is this? If we go— this is gonna show my home address. It doesn't matter, I trust you all. Look at that, we have people all over the world. We've got people from Indonesia. I'm from Australia. We've got people from Algeria. We've got people from the States. We've got people from Canada. A lot of people from India. Big shout out to everyone all over the world. Thank you so much for tuning in. It is truly a privilege to be speaking to you all. I appreciate it. But let's get out of that. Let's go back to the chat. We've got some more questions. South Africa, Nepal. Oh my gosh, I want to go to Nepal. I want to go to South Africa. I want to go to everything. Raza, I have just purchased TensorFlow for Udemy, TensorFlow for Deep Learning course on Udemy. Welcome to the course, Raza. Thank you so much for purchasing it. I hope you enjoy it. Let me know if there's any way I can help you. Philippines as well. This is beautiful. So Michael asked a question here. Do you have any experience? Haha, look at that. I love how it brought in the emoji. Michael, I kind of want to keep that there. I'm going to delete that because it's too big. I'm going to replace it with my own us. So Michael says, do I have any experience with or plans to teach ML streaming data? So no, no, I don't have experience with ML streaming data. However, this is definitely something I'm looking to learn more about. So I can't teach anything I don't have experience with. That's, that's just a rule. No skin in the game. But I'd encourage you to read this article, Michael. I think you might find it great. Chip Ewing, Machine Learning Is Going real time. So Chip Hewitt is one of my favorite machine learning role models. But the question you asked is more about streaming ML data. I personally don't have much experience with that, or hardly any experience. I'm more so of a static data type of person, which means data that lives in a database. So if you're watching and you're not sure the difference between streaming and static data, imagine static data is data that stays in a hard drive, doesn't change very often, whereas streaming data, you could imagine, is data that's constantly changing. So Imagine trying to live stream— or sorry, to try and transcribe this video as it's running. So that would be streaming data. But transcribing this video, so doing speech to text after the video had been recorded, that could be considered static data. So here we go. Great article. On streaming data and the future of ML systems. So really check out Chip Ewing's blog. Very, very good blog on machine learning and different stuff. And she's got a lot of great experience in different machine learning companies such as Snorkel, NVIDIA, Netflix, and, and many of the world's largest organizations develop and deploy machine learning systems. She also teaches CS329S which is going to be one of the places I would recommend you go after the course. I haven't fully finished the extensions after the course yet because I'm focusing on finishing the actual course yet, but I'm going to have a blog post soon of places I would recommend or just courses to look at and projects to think about after you finish the course, of course. Hello from Colombia. Hey Miguel, how you going, my friend? Stephen Walk says, where, when and where can I get you papers? I'm not sure what you mean by papers. Invalid. So Invalid asks, my friend never heard about TensorFlow, so how can I explain to him what TensorFlow is? Love from Bangladesh. Well, very big love from Australia, my friend. So in the simplest words that I can think of, TensorFlow is a programming library which helps with various machine learning requirements. So from turning data into numbers to finding patterns in those numbers, which is called modeling, to deploying your models to applications. Getting them in the hands of others. So TensorFlow is a full-blown machine learning framework. So if you have Python, which is a programming language— TensorFlow is actually in multiple versions, so it's not just Python, it's also JavaScript. So TensorFlow, the Python version, which is what the course is focused on— you have Python programming language, then you have TensorFlow, which is built in Python. to allow you to take some sort of dataset, find patterns in those data, or preprocess those data, which is modeling, and then get that machine learning model, uh, into the hands of others through some form of application. And it can help you with almost every step that you can think of in that entire chain, because what I just described is a very high-level overview. There may be a dozen steps, uh, throughout that whole thing, but TensorFlow is a programming language that's going to allow you, or programming framework or library, that's going to allow you to build those steps without you having to do everything from scratch. Let me know what you think of that definition of TensorFlow. If you have some experience with TensorFlow, maybe you could help our friend invalid out. That would be awesome. Chai China says Uh, Cho China, Cho China. Am I saying that right? Cho China. Hello from Thailand. Hello from Australia. Great to have you here. Um, so Nadir asks, Nadir asks, oh, what did I do? I didn't copy everything. I'm going to copy all of that. So Siobhan is right. What is happening here? Notion is not playing along with me. Siobhan is encouraging HTFCM to jump straight into Python. If you want to dabble in machine learning, go ahead, my friend. I have, I have 10 hours of TensorFlow video on my channel. If you know some Python, go and try this video out. If you find that you don't know what's going on in the video, there are plenty of resources to help you learn what's happening. But I tried to, as with all my videos, create them so that they're as beginner-friendly as possible. That's what my courses are. They're for beginners. I can't stress this enough. If you already have over a year experience in machine learning or TensorFlow, my courses aren't for you. They are for someone who is just getting into the field, who is wanting to try things out. And but we have this question. Here we go. Nadiru says, I've heard that AI will make web development will no longer exist. What do you think? So I have no predictions on the future of AI. These things are too hard to tell. So I have almost zero predictions on the future of AI, what will happen, right? Because every time you try to predict something that's like 10 years out, your prediction ends up being way off. So the further you try and predict out, the further, the more error is prone to your prediction. So who knows? Will I think web development will no longer exist? No, I think that will stay around for a very long time. Will it change? Will it augment? Of course it will. But that's the role. That's the nature of the tech field. If you get into the technology field, if you write code, any kind of code, you can basically say to yourself that anything I do today will probably be obsolete in, who knows, a year, 3 years, 5 years, right? So that's the nature of the tech field. Any type of software engineering or software field, that's the last 40 years. That's how things have gone, is things change almost every day. So I have no prediction of the future of AI. But I think web dev will be around for a very long time. Okay. Oh, we have a lot more questions here. Raza, here we go. Raza asks, what is the difference between your Complete Machine Learning and Data Science? They're both from ZTM. So, Raza, great question. So complete ML and data science equals absolute beginner, no prior knowledge of Python slash machine learning, and teaches you the fundamentals of MLDS. Whereas— I don't want this— whereas TensorFlow for deep learning equals follow on from above, builds upon Fundamental ML and DS skills with deep learning. So the Complete ML and Data Science Bootcamp, the ZTM course, is for people who have never— like, you can, you can start that course having never written any code before, and it will teach you the foundations of Python if you want to learn those. If you already know Python, you can jump straight into machine learning and data science. It will teach you the foundations such as NumPy, Matplotlib, Pandas. These are all libraries like Python libraries which are used to manipulate data. Whereas the TensorFlow for Deep Learning course is— it builds on the foundations so that you learn in the Complete ML and Data Science course, but it specifically focuses on deep learning, which is a relatively new paradigm. It's— I would say it's, it's almost about 10 years. Well, it's been around for a long time. However, It was in 2012 that it really started to ramp up. It specifically focuses on that using only TensorFlow, but we do borrow concepts from the Fundamentals course. So that's why the prerequisites for the TensorFlow course, if we come down here, prerequisites says 6 months writing Python code and at least one beginner machine learning course. However, if you are familiar with Python, I would encourage you to try out the TensorFlow for Deep Learning course. If it, if it sparks your curiosity, that is. Otherwise, start here and then go to here. All right. What do we have here? What's going on? Prateesh says, have you covered mathematics in this course? So not explicitly. So Prateesh, man, Notion, I would really love if you just highlighted this and let me turn it into bold. There we go. So not explicitly. The course has no straight up math lectures. We focus on writing TensorFlow code and teach the math concepts using code. So if you'd like to learn all the math that's going on behind this, I would recommend— so for the math behind the code we're writing, I'd recommend a course like— so if we have a look at this, deeplearning.ai, this is, this is probably the course that I'd recommend for learning the math behind what we're writing. Where is it? Programs, deep learning specialization. There we go. So that one. So that's very math-based. But our course is— or the TensorFlow course is very, very code-based. So if there's a concept that you should know, we teach it with code first. And that's encouraged. And that's actually the way I learned. So that's what I have the most experience with. So that's what it's taught with. So Weekly Hacker, great, great question. Thank you for pumping it in there. Does your course teach mathematics for beginners? We just answered that question, or sort of. So as above, the course focuses on code First, any math needed to know, we teach with code. Otherwise, extra resources are recommended for what's happening behind the scenes. If I'm intermediate in Python, should you buy this course? Yes, you would. Yep, you should try it out. You don't— you never have to buy the course, right? So the question here, should you buy this course or should you do this course? This is in the GitHub and the prerequisites are here, so you can decide for yourself whether or not you should buy that. I'm never going to tell you that you have to buy it. Um, okay. So I think that's— oh, Hellhound, thank you so much. I'm going to take a screenshot of that, uh, comment. I really appreciate that. Okay. So I asked, says, please make sure time series is covered very well. Other courses fail to do so. So the time series module is coming soon. So that will be released out shortly. All right. So as long as everyone is stoked, that is, we're going to wrap this up to prevent it from getting too long. But I think we, we answered a fair few questions. So it's very exciting. Of course, if you want the course, the best deal you're going to get, I'm going to link that into the chat now, is the Udemy TF launch. Uh, Atkash just joined in and said just signed up for the course. Very excited. Thank you, my friend. I hope you enjoy it. So the best deal is here. Um, if you want to lock it in the course, go for that. Uh, otherwise if you want to check out all the materials, be sure to check out, uh, the course GitHub. You can find everything there. Um, Atkash says, can I skip Food Vision and dive into NLP and time series directly? So, the course is built in sequential order, so I would encourage you to go through it sequentially. However, you, you don't necessarily have to do— I would encourage you to go up to everything 0 to 5 sequentially, and then you can go freely after that. The videos for NLP and skim lit are recorded. They're being uploaded as soon as possible, as I said in this question here, but everything else will be finished before the end of May. So time series and whatnot, they'll be coming as quickly as I can. Otherwise, thank you so much for joining in. I had a lot of fun. If you want to read all of the questions that we did, I'm going to share this. So share to the web. Um, there we go. Copy link. So I'll put this in the description as well. See the questions here. So there's all the questions and answers there. But thank you all so much. If you want to go and check out the course, where do you have to go? This is— I'm actually very, very excited that this is now launched. If you go to deburke.link/udemytflaunch, this will be the best deal that you're going to get. It'll last for the next few days or something like that to celebrate the launch. 30-day money-back guarantee if it's not up to your standard. There we go. TF2021 launch code is applied. So go and check that out. If you like this video, if you have another question, leave it in the comments. Otherwise, I will see you next time. We are— my 100% focus is going to be on finishing the course, but after that I have got some very, very, very exciting things in mind for the future. So keep learning, keep moving. Remember, if you're coding Python, you've got to be rocking pythons, and I'll see you next time. Thanks for joining in. Peace.