LEONARDO.AI Tutorial - How to use this fantastic Ai Art Platform!

if you're looking for an AI art platform that will rival mid-journer's quality with the ease of use then check out leonardo.ai this video I'm going to go through and show you just how powerful Leonardo AI is it is free you do get a certain number of generations but it's actually a lot more than most of the others and the quality is absolutely insane the other thing is at the moment you do need to actually get whitelisted so you need to just pop your email in here and uh you'll get white eventually I'll get back to you whitelist you and you can get access to the platform now before we start these are some of the community images check out this woman's face with a soft skin the detail is great the picture is crisp and clean really amazing uh artwork generated there we've got this cyberpunk girl if I zoom in just how these images are so exceptionally clean and then you get something like this where you have a lot of detail but apart from a few glitches here this is like just absolutely amazing and check out this same absolutely incredible it's so clean so precise and it just has really great composition the way it's laid out also we have this kind of owl that's made out of jewelry just incredible quality we're going to cut through a little bit quickly here but I'm going to skip past these and come back to them but we do have here AI canvas and image AI image generation first of all the canvas it's essentially somewhere where you can generate images and then add to them we're going to explore that in the AI image generation which is a simple prompt based interface with a few options down the left here allowing you to create some pretty cool images what's really cool is you've got your community feed here much like mid-journey you can see what's been created recently and get some inspiration for prompts and see what models have been used you've got your personal feed I haven't created anything yet because I've just signed up for this account I've tested it out on another account but I've signed up with this one and we've got training and data sets so you can actually create new data sets if you want to look into that and start to create your own AI models based on images you use I haven't actually played with this at all but I might check it out in a future video and you can check out the fine-tuned models these are models that are on Leonardo AI so if you're creating a particular type of image maybe it's pixel art you can choose the pixel art model if you're trying to create a potion use the potions model there's all these different models for very specific uses including these Christmas stickers Crystal deposits and these are sort of like uh models made by the platform but there are also Community models so if I go in here to community models the community have actually created some pretty cool and unique models themselves and you can use these for your AI art generation this is where Leonardo AI I think is probably the most complete and exciting of the AI art platform so far because it not only has incredible capability but all these different options and features and models you can use it's definitely a future Contender but let's go down to AI image generation and see how this works first of all I can choose how many images I want one two four eight for the sake of this I'm actually going to say eight I've got here Skeletor playing an electric guitar it's actually saved that from my last session which from a different account we'll stick with that for now although we might just say skeleton playing an electric guitar I can choose the resolution I'm actually going to go 512 by 512 just so it generates a bit quicker but I can choose what resolution I want to run at this may change depending on the model I choose and I can also adjust here if I want it to be a different width so 640 by 512 making it four to three I can choose my aspect ratio and it will adjust accordingly uh so that's pretty cool the guided scale if you've used tablet Fusion essentially this is how close or accurate it is to what you're putting in too far down is going to produce poor results and if you go too far it's going to adhere exactly to your results but uh somewhere in the middle is pretty good uh I find I'm just going to stick with the default settings for this the step count is actually if I use one or two it'll only produce one or two steps or iterations of the image which is not going to be great if I put it up to 30 generally you'll get a cohesive image and further than that it will continue to add to the image but it may not actually add detail it may just simply change the image a little bit but what's also cool is the tiling feature if I put tile on here I can create an image that tiles like a pattern and go from there which is pretty cool so that way if your graphics design you're going to create a pretty cool pattern you can do that as well I also have imaged image and if you're familiar with the idea of seeds you can choose a seed number and use a fixed seed so that every time you generate an image you're starting with the same fan foundation so you would essentially be getting the same images if you use the same prompt but we're going to keep pretty much all this at default and we'll explore some of these elements as we go so I've got a skeleton with green glowing eyes playing an electric guitar I can choose my model here so it can be isometric sci-fi buildings or Leonardo creative some of these are actually Community models I'm going to stick with Leonardo's signature you can choose a style Leonardo style or none we'll just stick with Leonardo star because that's I think what really makes these sort of really good with the negative props I can say no cape so you've got the option for a negative prompt your positive prompt I'm going to generate so I have two tokens and I have 250 for the day and it resets in 20 hours so let's generate and see how we go so now we have our images if I click on these I go through this one it looks like uh some headphones or something it's a little bit strange there's other ones with the guitar this one here is probably the best so far or this one here so pretty cool very neon look what I can do is I like this one here I can download it pretty much as it is I can unzoom the image I have no idea what that does let's try it out so you see here I have an unzoomed image and I have the original image so unzoomed actually brings it out so if you're creating an image and it's cutting off various areas you can unzoom that image and it will add information around it and also here I can remove the background if I want to so let's actually give that a crack as well so now I have an image with no background I've opened it up in Photoshop and I actually have a transparent PNG so you've got some pretty cool options here we've got edit in canvas if I want to as well as upscale so I'm going to do is I'm going to upscale this image to get the full quality version and check out the difference in quality it's really taken that next step and what's really cool about this is our original image was 896 by 512 but this upscaled version if I take it into Photoshop this image is 3136 pixels wide now getting that at a mid Journey would be a bit of a tall order so you can see we create vastly higher resolution images in Leonardo Ai and that upscale added to the Quality very well and actually looks pretty good it looks like a really solid generation but let's move on let's bring this back to two images to save on some of our tokens and let's try a few other models to see what we get let's go from Leonardo's signature to Leonardo select and generate you can see now unlike last time it took the grand applied to the skeleton I did say green glowing ice so it's good to see the eyes glowing green so these are both pretty cool images one thing you may notice though is that actually Square because when I changed the model it actually changed the resolution to fit that model what I'm actually going to do now is I'm going to upscale this image and take a look and check out that it's added a bit more detail and it's made it a pretty fun looking image which I think is pretty cool now let's try that same prompt again this time on Leonardo creative and again we get something different once more once again we'll upscale to see how the final product looks and again we've got something that is pretty cool but why don't we start exploring some of the other features because we've tried the three Leonardo platforms here there's creative select and signature we can just use straight UPS stable diffusion but we'll stick with Leonardo for now because that way we can see what Leonardo has to offer in regards to its own unique platform I'm going to go down here to tile and I'm going to turn on tile and then I'm going to find a different model to try and create some kind of pattern so I can go down here and choose select custom model we're going to generate with this Crystal deposits model we're going to remove our negative prompt and try something a little different now this is not what that model is probably designed for but I already give it a go and see what it comes up with you see here I've now got these sort of patterns there's no face involved but that's okay because that's not the model we're using but I'm going to take it back into Photoshop and you can see that it has tiled pretty seamlessly so that tiling feature could be pretty handy if you want to play around with some certain features and Leonardo AI to get a particular sort of pattern or effect that you want to repeat on a design or on a project could be really good for use in video game textures and things like that so now let's try something different let's head down here image to image I've uploaded this alien image that I created many years ago I can choose pose to image which means that it will try and use the image to kind of pose what it is to generate so I'm going to turn that on I'm going to scroll up I'm going to say now we're going to go with a different model because Crystal deposits doesn't really suit let's go back to we can try a custom model or a community model let's try imaginary friends and generate with that model we've still got our image poster image taking a selfie so I'm going to hit generate and it should produce it in this style that we've chosen so you can see it's kind of kept the same basic sort of shape in there obviously it's not perfect but it still looks pretty cool and I think if you were to upscale this it looked pretty good but we've only got so many credits so I'm not going to upscale this one in particular but uh you can see how that works and I can also then try switching back to Leonardo's signature and it's kind of done something funny with the face over here but uh this one looks a bit cooler but you can see even though it has messed up and repeated the face the detail in the face is really quite really quite good and we're gonna go here to prompt generation so it will actually produce some prompt ideas for us so let's say powerful Warrior and hit id8 so I've got some pretty cool ideas here a fierce Warrior staying on top of the mountain of skulls brandishing a flaming sword probably doing Shining Armor surrounded by a swirling storm of energy let's try that one okay generate so you can generate prompts themselves as well as the artwork to have this Dr Strange look like and we've got this child here so uh pretty cool looking images um not exactly what I had in mind but still pretty cool so I'm going to try now I'm actually going to go back to the community Tab and I'm going to try and find something as a bit of inspiration I really like this hyper realistic style so you see there's a lot more sort of detail in the prompt he has used image to image I believe but possibly by looks of it but I'm actually just going to copy this prompt this used a fine-tuned model RPG 4.0 so let's see what results we can get by actually playing with that prompt so I'm gonna go back to our AI image generation I'm going to choose a custom model RPG 4.0 which it looks pretty good we paste in our prompt and now we're going to change a few things now this person's put a lot more effort to their prompts because they're actually trying to generate art we're just basically showing how this works I'm gonna put a cyberpunk human wearing [Music] red sun glasses like undercut hairs I'm going to say dreadlocks headphones The Cinematic light award-winning Masterpiece let's just stick with what we have there and generate that you can see now we've got some pretty pretty nifty results so if I upscale this image because it looks looks pretty good looks like a pretty uh solid picture I'm going to hit upscale and check out our upscale image the hair has got some great details the eyes are looking a little drawn on but otherwise like the face is looking pretty good for what it is it's actually it's still a very high quality render and something worth having to play with and remember if you want to play with this further check out those custom models got a bunch of platform models here which look really good really uh eye-catching very specific use case in some of these such as character portraits uh sci-fi buildings there's so many different styles you can play with especially when you start to go through the community models and have a look there you can get some pretty cool stuff out of that as well now one thing we can also do is I can go into this here and I can edit in canvas if I click that I get this canvas mode so that means I can actually move this change my width to the same width over here and I can add to this generation so I can say Orange dreadlocks and generate and before I actually choose which one I want to go with I can actually scroll through see which one's going to suit best the first one is definitely the best of those I click accept and I can then go through and do other things I have draw mask so if I want to draw and change my brush size draw a mask over here I say green sunglasses on a person with orange dreadlocks and I do believe if I just say green sunglasses on a person with green dreadlocks I'm not sure where this will change the sunglasses or the hair so let's just see which one it affects as you can see it's changed the sunglasses but left the rest of it unchanged but I can also go in here to erase and this is actually just masking only or I can erase so I want to erase the background I could do that or mask only I need to actually sorry I need to scroll through and decide which one of these I want to keep the second one except and now I can use my erase tool to erase some of this as well it's time I hit right lights hit generate and you see it's filled in those areas over rest I've got all the usual options over here I can choose which models I want to actually use on here because I've generated this with a different model it's cool I can use the image prompt generator to create something come in here and then just adjust it with stable diffusion and I have all the usual stable diffusion options here but if I'm done with this I'm going to say exit the editor for now if I decide I can go back into AI canvas I don't want to work in this image I can just click on it hit delete and I can actually upload my own image to work with now I can choose a I'm gonna shrink this down because it is a big image but I can choose a JPEG or in this case I've actually chosen a transparent PNG and I want to change this eye so I'm actually going to just pop the viewport here and I can actually make changes to my image I can go here to draw a mask make my brush nice and big and you can see there's an ant on here I can actually probably add to that a little while too so maybe I'll just say draw some spots here and I can say a hand with a purple eye on it ants crawling and hit generate and now it's out of the eye it's also added a background I can choose various Generations as usual I'm going to stick with this one except and it's actually changed my image and I can do the same thing if I want to I can bring this down now I want to try and get that to snap because I want to add a background so we say here now a wrist against a background of dirt and lava hit generate and now I have these different Generations I can choose from I'm going to stick with number three this time I'm going to switch to stable to Fusion 2.1 I'm going to come down I'm also going to keep this here but I'm going to say a human face I guess the background of dirt I don't know how it's going to turn out but I'm also going to draw a mask here hoping that it puts the face on there I actually don't know what it's going to do but let's hit generate and see there is a face on there doesn't look great this one here is probably the best I hit accept so I can have some fun actually just adding to this image and changing it using AI so I've actually combined my artwork with AI to make changes to my own images and you can do this photos artwork um pretty much anything you want and of course if you want to you can upload an image from previous generations or from the community so if I want to play with this image I can also pop it in here and then have a play with that I can make some changes to that if I want to I've put closed mouth but it's produced an open mouth so you know whatever works when I'm done I can download the artwork and you can see over here in Photoshop I have this large PNG with multiple images on it so I can do what I want to with that but it's a pretty cool feature so you see how by playing with leonardo.ai you can produce some amazing imagery uh it takes a little bit to learn like any platform I've just been having a bit of a play at it I've had a little bit of a play but nothing too crazy but um the artwork you can produce really is next level I believe it is probably the main Contender to bid Journey at the moment it has the quality close to the quality of mid Journey without having to be too technical with your with your props but also has the ease of interface that doll that Dolly has so it's kind of like a nice blend of the two very powerful tool and I highly recommend you check it out there's a link in the description leonardo.ai get yourself whitelisted and it might take a week or so and then get in there have a play ton of fun uh I'll be having a play right now but otherwise if you enjoyed the video please consider giving it a like otherwise I hope to see you again next time have a great day

Channel: Wade McMaster

Views: 155,531

Keywords: leonardo ai, leonardoai, leonardo ai tutorial, ai art, ai art generator, artificial intelligence

Id: 36rSjS5hV4Y

Length: 19min 9sec (1149 seconds)

Published: Tue Mar 07 2023