 week. Thank you all for stopping by on YouTube. So, I'll stall. I am definitely... It looks like I can see a little bit of activity happening for thoughts instead. Yeah, you can just do a sort of a play along where you play the character of me trying to get this stream going. It is just buffering, huh? All right, let's see. Stream setting, stream quality. Yeah, boy, I'll tell you what, I'm not broadcasting at a high bit depth either. Let me double check some output settings here. It's okay. We will do this. Yeah, this is going out at 720 to our R&TP server. Yeah, all that looks good. It's not super high. It's an 1878k bit rate. It's okay now. Now you're okay. Now it looks good, huh? All right. I will let it cook for a little bit. Let it simmer before we get into the meat of things. How about that? Looks like we're working. All right. So, maybe we should start the show early and let it work all its kinks out in the future. But for now, I'm going to go with it. And since it seems to be working, what I want to do is kick it off and say, hey, welcome to the show. And before things go haywire, I want you to head over to this page, which is where our product pick of the week is happening this week. Head to that page because there you're going to get a 50% discount during the live stream. And the page looks an awful lot like this, so there's a hint as to what the product pick is. So head on over there and you're going to find a great big discount. In fact, if I go and refresh right now across our fingers, we should see boom, half off on this product pick of the week. So why don't we now jump over to Lady Aida and tell us a little bit about what we've got in store today. Take it away. Raincraft Hat is our idea of how you want to make machine learning projects for boards like the Raspberry Pi. So we wanted to make something that makes it easy to do machine learning projects. So I actually went and built a bunch of projects using TensorFlow, using Google Assistant, using Alexa. And it's like, OK, well, what did I learn from all these? What are the things that would have been really handy while building these projects? For example, TensorFlow Lite runs on a Raspberry Pi 4. So there's a Pi 4 underneath here. So to do vision projects, one of the things I noticed is that it was really annoying to do them without having a visual representation of what's going on. So in order for people to make easy projects, I decided, OK, let's put a little TFT display. So the display shows exactly what's on the camera, what the camera is seeing. Hopefully you'll see like human or something. No, I'm not human. So it's got a display 240 by 240 display for showing what is on the camera. I can also show you the frames per second. I've got the fan off, so that's why it's telling me the power is dipping a little bit. So having the display is really helpful, so you can do vision projects. There's a slot here for a camera, so you can use the Raspberry Pi camera with it. There's three dot star LEDs. There's a three-way joystick, and there's a button, so you can use that for audio projects. We'll show that. There's two JST connectors, so you can connect servos or relays or NeoPixels. That's something that actually the folks at Google TensorFlow said would be awesome. They're like, I want to have a way for you to do machine learning, and then it controls a relay or servo or motor in some way, so we made it so it's plug and play. You can plug in here or plug in here to connect another device. There's two speaker ports. Tiger. It talks. It talks. So it's just a tiger. There's also a headphone output and two microphones, so you can do audio recognition projects. And most importantly, there's an on-off switch for the audio. So if you want to ever make it so it's definitely not listening, this will manually disconnect the power from the audio circuitry. So it's definitely not listening to you. One thing that we noticed is when you're doing especially video recognition projects, the Raspberry Pi can start to overheat. So we added a fan that attaches onto it, and then when you plug it in, it turns on the fan, and it will cool down the Raspberry Pi so you can run your projects longer. We definitely noticed when we were running it for 15 minutes, it wasn't a big deal, but when you run it for an hour, the Pi slowly heats up. So there's a fan that comes with it as well. It works with our existing TensorFlow Lite project guide that we published. And we also published a guide on how to use a teachable machine from Google to create a TensorFlow Lite file that you can then run on the Raspberry Pi to recognize custom objects. So yeah, I think it's kind of cool. I think this will make it easier for people to start exploring TensorFlow Lite and other machine learning projects. Like I saw today, there was a project somebody did with a Pi that recognized license plates. And I think when you have a standalone computer that can do some of these projects, they might take off. Yes, that's right. That is our product pick of the week. And I'm going to go head over to my cabinet of mystery drawers and grab one. I'll be right back. Yes, that is right. That's our product pick of the week this week. It is the BrainCraft hat. This is a hat for the Raspberry Pi, and it works particularly well with the Raspberry Pi 4 if you were doing machine learning TensorFlow Lite types of projects. And in order to support that, we've got on here the TFT display. It's a 240 by 240. We have headphone out. We have speaker outs. There's a joystick with a push button on it, as well as a second button. There's some, not neopixels, there's some dot stars for giving you some feedback on your projects. There are two microphones on there, which I believe you can use for things like, A, just getting wider coverage, but also you might be able to tell what object is closer based on the sort of stereo of the two microphones on there. Don't quote me on that, but I think there's some possibilities there. What else does it have? The StemEQ-T connector on there is great for adding any of these types of sensors to your projects using Blinka libraries inside of Python, as well as writing out to things like relays and neopixel strips over the 3-pin JST connectors. So a lot of connectivity there. You can use it for all kinds of projects, not just machine learning, but what I wanted to do is demo a little bit of the machine learning projects that we have. First thing is, check out, if you head over to our page for this, this is product ID 4374. So you can head to that URL, adafruit.it, slash 4374. You'll see it's half off right now during this broadcast, and you'll see the broadcast is actually happening right inside of the product page right there. If you scroll down, you'll get a bunch of info about the BrainCraft hat, and you'll get down to some links to some learn guides. So you can go to the main learn guide for the BrainCraft hat that helps you get started with it, get all the proper libraries and things installed on your Raspberry Pi to get up and running. Once you do that, you can kind of move on to this running TensorFlow Lite object recognition on the Raspberry Pi 4 learn guide, which I went through just recently to get this set up. And note, it takes a little time. You want to carve out some time to get all of these libraries and things installed, all these tools installed on your Raspberry Pi, but once you have it up and running, it's really pretty darn cool. So what I wanted to do is actually show a little demo. Let me pop over to, actually, let's do a full screen. There's a full screen of my Raspberry Pi 4 with the BrainCraft hat right on top of it. One thing you'll notice, actually, if I touch the second field, it's vibrating a little bit, and that's because it's been running for a while, and so the little fan that's built on there is spinning in order to cool it down. And what you'll see here over in my shell, over in this terminal, is I'm running the TensorFlow Lite object recognition, and it is searching for objects. Currently, you'll see it's mostly saying, none, it hasn't found that much. It might be seeing this little line here and trying to fit that to some things that it knows about. And I actually have the headphone output going into a little amplifier and speaker, so you should be able to hear things as I show them. So let's start off with a really popular one. Oh, it's not doing the audio thing. Hold on, let me fix that. Let's try that again. So if you take an object away and put it back in, it will rerun and it will play the audio again. Coffee mug. Coffee mug. Excellent, it did a good job with that. How about it nailed it. Screwdriver, okay, let's keep trying. Granny Smith, I believe. Ping Pong Ball. It thinks that's a ping pong ball. Also, somewhat surprisingly, I have an overly ripe banana here, and it doesn't know what to do with that. You can look actually, if you take a look at the text that's going by, what's it trying? Slug. Banana. Oh, it did. Vinesnake. All right, it's trying. It's trying some things there. So I'm going to give up on the banana. And it's a whole lot of fun to try things on this. And of course, one of the things with machine learning is that if you have a type of category of objects that you want to use that haven't been trained into the system, you can train them. So you can show it a lot of examples, and they can be photographs. It doesn't have to be the actual objects using the camera the way this does. Let's see. I don't think it figured this one out either. Here's a Crescent wrench. Adjustable wrench. I think it tried can opener. It was about the closest that it got, depending. It thinks it's a can opener. The last one I want to try actually is, just for the heck of it, what happens with Lars? What does it think of Lars? It thinks Lars is a teddy. That's very cute. He's not, and he's not safe. So don't use that as an example of things you should do, kids. So that's a little bit of a demo of the BrainCraft hat in action. There it is. It's really a complete all-in-one board for adding a lot of I.O. and capabilities to your Raspberry Pi. You can do a lot with it beyond just what it was designed for, which is this machine learning. But it's particularly great for that. You'll notice also we have a little slot there for the ribbon cable to go in. I can show you my full setup here. I will point the camera at itself, and you'll see there's my ribbon cable coming. Oh, hey, it found the banana. There is the Raspberry Pi with the BrainCraft hat on top of it. And if I lift this all up, some things will probably come apart, but I'll show you in the main camera view here. Hold on, I'm going to switch my camera view back to me. You'll see that's all it is. So I've got the Raspberry Pi. I'm running it off of Ethernet instead of Wi-Fi. It actually doesn't need a connection at all at this point other than I SSH'd into it to get it started. And there I have one of the little Raspberry Pi cameras on a small little bendy tripod and power, and that's about it. I believe the high-quality camera would work as well, which the clarity of the image is probably helpful. So you could try one of these high-quality Raspberry Pi cameras with a different lens on there to see how that goes. But that is the thrust of what you can do as well as a whole lot more, as Lady Aida was saying, some people have trained it up on license plates. And I think this type of object recognition and machine learning on the Raspberry Pi is a really powerful and sort of open field. So you can really learn a lot and try some things out on that. And this will get you started. And the kind of amazing thing is it's going to do it for a low, low price right now during this live stream of $22.48, which is tremendous. Maximum of 10 per customer. So if you have a big machine project plan maybe coming in the fall semester here in a school or university setting, maybe get hooked up right now. But here's the thing, that insane price is only good right now during this live stream. So you want to hop on there, load up your cart and go, and then share with us what you do. We're excited to see what types of projects people are doing with machine learning. And I'll be hanging out in the chat afterwards if you want to chat about some of the things you're doing or any questions that you might have. And so I think that's going to do it. So thank you all so much for stopping by. That is the product pick of the week this week. It is the BrainCraft hat for machine learning on the Raspberry Pi. And I will go ahead and hang that with pride on my product pick of the week board here. And that's going to do it. I will see you next time. Thanks everyone for stopping by. Take care.