Previously, Ada taught many applications of AI drawing, Such as Stable Diffusion and Midjourney, However, they each have their own limitations, For example, With Stable Diffusion, your computer hardware can't be too outdated, In principle, if you don't have a dedicated GPU, It's not going to run very smoothly, As for Midjourney and Niji Journey, Although they produce beautiful drawings, And it doesn't require any specific computer hardware threshold. However, it costs money. And a certain level of language ability is needed. That is, you must use English commands. For friends who are not good at English, it can be somewhat difficult. Although Open AI has recently made the AI drawing tool DALL-E3 available, it functions quite well. But if you haven't subscribed to ChatGPT Plus, you won't be able to use it smoothly. So today, Ada is here to share a few useful ones, Let's learn it together The first one is the online version of Stable Diffusion Its method of use is very simple As long as everyone enters this URL below And then you can see he has a 'Get Start Now' here Then we click into it In principle Its usage has no restrictions If you let it develop slowly It can keep going on and on and on You can make N slides Its advantages That is, the images he produced are bigger And he is using a model of SDXL So theoretically, the figure he made It will also be a bit more beautiful Next, we will begin to input the spell Because there are many spells If you're talking about writing spells (narratives) If you have no concept, You can go to websites like Prompthero to find here There are many examples inside Then everyone can see It has Stable Diffusion on top here Then there is Midjourney With ChatGPT, Various syntaxes are available If you're not sure what spell to curse with You can check it out here Then you can see He shows some examples here Let's find one first Like this A pretty good sexy picture Look at this picture, Beside it will be a spell in both positive and negative directions We just copied directly Come here. Here's a prompt code, you can just paste it directly And then down here It does have a downside Let's open it This is in reverse We can duplicate it the same way There's a downside to this service actually. Meaning its model It only has SDXL So what it can draw is less rich You all can take a look, About this default value. This has the best image quality And then the sai below Actually, it's just something he calls 'style' Of course, I can still draw Everyone can give it a try Does it resemble what you want to draw? After we have entered both positive and negative words here We will directly select this one to generate In principle, You don't need to join his premium membership you can draw without any limits. But the drawing speed won't be that fast If your computer level is bad Maybe it's not even faster than it It takes him about 50 seconds to process one image Everyone can see the timer counting down below Then wait a bit It is calculated The image this website creates There is a little problem, that is, He only shows the upper half of the image If you want to view the whole picture, He has a download option here You just need to click once Then open it here and take a look The picture is now out If it's a free version, it can only function at 1024*1024 Don't think it's small Because the normal Stable Diffusion can only do 512*512, Unless you specifically adjust the size So, this is a free service The first limitation is If you are using the free version The speed at which it calculates the picture may be slower Then the second one There are fewer model options What about the third one? The size of the created image is fixed You can adjust it up to 1024*1024 at most on your side The fourth one is that his picture has a watermark in the lower right corner Didn't we just say that he can choose the 'style'? This is the movie mode We can also choose this Then we can choose this comic book He can redraw it in comic mode Let's look at the comic mode Everyone, take a look That's how it is produced He is just a model of a cartoon image Basically, he draws based on the SDXL base model. there's less choice Next, let's introduce the second one for everyone I think a better service Next, let me introduce the second one It's the same kind of Stable Diffusion service It's called Tensor.Art His website is also here Everyone can input on their own First, we need to log in He has a login at the top right corner You can use accounts like Google or Discord Or you can use an email account Then I can simply use Google to log in on my end Because I've already registered So, it directly allowed me to log in So the feature of this website is that His usage is almost the same as Stable Diffusion But it has more features It supports WenSheng diagrams, and TuSheng diagrams You can even generate GIFs You can also use VAE, LORA, ControlNet, etc. And there are a lot of templates Next, I will implement it directly for everyone to see First, we can see here Its first page option is called Model There is a lot on this side You can see there are many models available What about here? It has some sort of animation or concept Just find one Let's choose something more human-like Many people have been using this recently This is called simplemix V3 Assume you like this model Just move your mouse over it Then click on this run Next, another window will pop up Over here He asked if you want to use his recommended parameters It's best for everyone to use it directly Because if not used Its picture might look a bit strange. Is like our previous VAE without a sleeve The color is either light or weird In principle, you should use the default Then everyone can see here It also has a positive prompt word, Follow the reverse prompts The reverse prompt word has a preset You don't have to fill it in. Or, as we have in our prompt hero Then we can directly fill in This place in the upper left corner This is called the WenSheng diagram This is the image that I created I've tested and made a few They're all quite beautiful The image below is a raw image I mentioned before making the chart You can send an anime picture and then convert it to a real person The idea of 'Wen Sheng Dong Tu' is you can choose a model, Then enter some text to generate the animation I will give a simple demonstration here We chose this one just the same Use this as an example But because the models differ in length So the picture he draws will definitely be different Let's apply a positive tag to it We can leave the negative term blank on our end Look down here Everyone can see here, It has all proportions You can also adjust the size yourself The advantage of this website is that It can probably create about 100 images a day without any problem If you finish the drawing today Come again tomorrow Actually, when speaking seriously, it's sufficient There are many options available under it For example, the CFG we mentioned before That seed It even supports ControlNet I think it's not bad As for the sampling algorithm here Everyone can choose by themselves Ada had mentioned it before If you want it to resemble a human You can choose similar ones like these These few are all good The default value is also possible If you're using the free version, He can draw two pictures at the same time at most We're going to draw two pictures this time You have to pay for more than 4 We'll just click on this generate online Then it has another weakness That is, he only speaks English, Simplified Chinese, and other languages Everyone just make do with it. Or you can use English as well It's quite beautiful Of course. The disadvantage of Stable Diffusion has come out Did anyone notice this corner deformed? If you like it, you can download it directly If this one is good You just directly click on this You can download it It's just that simple But the original picture is relatively small So you can magnify it by using other applications One advantage it has is It has many models So you could say that this model Don't want When you click in like this, Let me try again Let me show you guys This is his model just click once more Did you see it? There're a bunch of models you can choose from here I suggest everyone select 'all' In this way, you can view more types, He has a lot of models And you can even download the model yourself Then upload it for use If there are no models you like here So this is really, really powerful Now I'm going to switch to using an anime I use this in an unchanging state Then let it generate, The animation tends to have fewer mistakes This service is very good There are no restrictions on this service when it comes to drawing So, if you want to create that colorful picture, That's also possible But don't ask me how to do it Everyone can come here and find it on their own Or use some of your own imagination It's very easy, The animated picture is the same Also this Then you can choose a basic model first Assume there are many models here It depends on which one you want to choose Let's pick one at random Let's first find 'all' then Then, as usual, enter this prompt And then it's generated online That GF uses more computing power It deducts 12 points at once He probably provides about 96 computing powers each day So everyone needs to consider a bit The image will be moving Isn't it impressive? But it's not for making videos He is making a GIF So, if you want to make an animation You might just download this picture And then take it to something like Runway, Throw it over there for animation, Like this is also possible There are quite a few very talented people doing this on the Internet nowadays So this is the second thing we're introducing to everyone Tensor.Art This is quite good This is my personal recommendation If we were to mention a drawback It's just that the functions are a bit too much Therefore, you need to research slowly We'll basically conclude here, Next, let's discuss the next one The third one Introducing this to everyone Named YODAYO It's also the same, a type of online service called Stable Diffusion He gives you the same free points every day When you need to use it here Just click on this Login at the top right Then you can simply register using that account It is quite simple We just do it like this Then it's in It's very convenient When you come in here, Everyone can see there are many examples made by others here you can refer to these, He has some examples on his side he even gives you the syntax, If you say you can't use it You want to draw All you have to do is copy it down and paste it Drawing pictures is also quite simple We just need to click this AI art, We've imported it And then over here You can still input both forward and reverse words. We input the positive word as always Open it up It has a default setting here, you can choose to use it or not In principle, you can use his default If you say you don't want to use it You can directly modify like this Put on your own All these are possible Okay, and after it is done, He can draw directly You can see It gives you a number of free beans every day I guess it's about 150 points I didn't make a note of this. Then draw a picture with five points, However As everyone can see If it has a high priority If you shut it off, The beans turned into zero There's no limit to how many you can draw But it will take a very, very long time If you have nothing to do You can use this zero point Now I will directly give it 5 points This is done very quickly It's pretty beautiful Then he can also choose the model Everyone can see Its default is the epiC photo model Then we can directly switch to others Let's switch to this one do you see? It's very beautiful You can download and use it This drawing is also quite beautiful In principle, the drawings are based on the model you choose Make a choice As for what pose it adopts If you really run out of ideas Adam, I am also telling everyone here There are so many prompt heroes here You can search for some reference words to use Just drew this many I'll just randomly select another one Like this picture, it's done very well In principle, it depends on which model you choose The better the model you choose, the better it looks The speed of making this chart is also very, very fast Then this thing Its features are not as many as the features we introduced earlier But the basics are like literature illustrations Or a static image like this one can be used too You can directly upload images here Let him modify the model Essentially, everything is feasible As for how to do it Everyone should research on their own It also has a ControlNet But due to the length of time I won't talk too much Go ahead and play around What's next is Let me introduce another one to everyone. I think it is a software that can replace something like 'Midjourney' This is DomoAI we have a tutorial on our website It can help you turn photos and videos into a GIF Or turn it into animation Like this This is a short clip from when I went to Harry Potter It would help me create a matrix-style video it's very simple How should this website be used? First, you need to go to a homepage on Discord Come to our webpage You just need to click into it After we entered Discord First, you need to log into my account After exploring here You need to input this DomoAI And then click on it You can use it just by adding this you can see here at DomoAI There are some things on his left half Generally Normally generated, it can make images And what's below This can be made into a video You can upload the video directly after this Then just throw it up directly It will do it for you There's a NSFW below If you are not 18 yet Never click to look inside I can only say that much Let me first introduce to you all how to generate images Let's first draw this diagonal line And hit GEN You can issue a command here For example, I will give a case about one girl Then smile, Good! Everyone, look The image from this has come out It's that simple In principle, what he made here It can also be enlarged U1, U2, U3, U4 Let it make other changes Of course, the more detailed you write The more functions it has This is anime-styled Let's take this as an example Let me go ahead and make a chart Then let's directly test it It also supports features such as turning anime into real people. I'm going to use it directly now I did choose other models just now I finally made this figure Then we can just Let's save this image for now Next, let me briefly explain for everyone In this service the same Let's say you have a cartoon image How to make it look like a real person We just input the command /real here And then you can upload pictures next. Let's upload this picture It's out Everyone can look This picture This on the right is what I uploaded to him This is an animation picture This is the real person image it calculated What does everyone think? And the last one? The function is to make videos Here we can choose the option to generate a video So, we can type like this in this corner/video, Enter At this moment Then he can upload a clip to him Next, we'll simply upload a video for it After uploading This will allow you to choose which mode to use Then we directly use this to generate five seconds use the original mode, And then use the original video directly Then we'll just use this Then we can use the animated model for the model. We will use the Japanese anime model Finally came out Animation is indeed slower I can directly check here What it looks like when he does it Honestly speaking, it's not bad at all However, the duration is probably only about 5 seconds So to sum up this DomoAI it has several features It can draw anime pictures like Midjourney And all you have to do is modify the model It can draw portraits just like Midjourney So it can also help you get the anime picture Make it seem like a real person The other best feature is that It can help you turn real-life videos into cartoon videos Currently, all of these services are free of charge This is DomoAI Next, finally Let me introduce Copilot to everyone Or an AI drawing from Bing Moving on to the last one Previously, we introduced BING AI drawing What about before? We are on a website of Bing Then it has a drawing function Now, due to the updates to the browser and Windows 11 It will actually produce a Copiolt Right now, I recommend everyone to use Copiolt for drawing Because using Bing will deduct points But if you use Copilot to draw, It doesn't deduct points So you can draw without restrictions This tip is actually a secret trick Everyone please use it discreetly How should it be used? Technically, on the browser side We just need to point out, Copiolt will pop up Next, we can input directly here Help me draw Pictures Then afterwards, you just press Shift plus Enter Then we fill in the description The advantage of Bing's input is Its understanding of Chinese is very strong, So, if you have any ideas You just type it into Bing He can draw it directly for you Let's start with a simple example First enter a section Red background. Just use simple golden lines to draw a Chinese dragon Make it match the occasion Because the New Year is coming soon Then we'll send it directly Generally, it will display 1 to 4 pictures it's usually four cards, Then you can just open it directly It would then take you to Bing for viewing See which one you like You can just download and use it directly It's just that simple If you're drawing normally Actually, you come here and order this And entering here is also the same But it consumes points Over here at Copiolt You don't need to deduct points for inputting here If there's no catch I recommend it to everyone Many FB groups are teaching how to do AI drawing This is the vocalist of DALL-E 3 now All you need to do is look There will be some examples below We can directly use the examples we like Copy it down Paste it directly. There's an example here that allows you to modify, We dress him in an apron I suggest you talk to Copiolt You can add a 'please help me draw a picture' at the very front Otherwise, he might not understand what you want to do Did you see it? Quite simple Quite adorable It doesn't deduct any points for this either Its main disadvantage is its character representation It feels like plastic No matter how you draw You might think it doesn't look like a real person If you really want it to look like a real person You must use Midjourney And then draw with the V6 model It will be very close to a real person The second disadvantage is, How he draws his pictures are all squares Like some, we can request to draw rectangles or proportional diagrams He couldn't No matter how he draws, it's always a square But his power is Even can input Tao Te Ching, Buddhist scriptures, Tang poems He can draw pictures that fit the mood very well As mentioned above These are the free AI drawing services introduced by Ada. Except for the official Stable Diffusion, which is a bit weak Honestly speaking, the rest is pretty good So, if you guys don't have a catch You can go to prompt hero Or go to the few websites I recommended I always put relevant resources in the description box, It's for everyone to refer to, I am the Computer King Ada If you like this kind of teaching Remember to like, subscribe, and turn on notifications for me See you in our next video! Goodbye
