Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
1
00:00:00,230 --> 00:00:07,670
Let's take a quick look what we do over the next lectures to make sure that we have a good AI influencer
2
00:00:07,670 --> 00:00:11,210
that looks similar in all pictures.
3
00:00:11,330 --> 00:00:16,070
So first things first we use stable diffusion.
4
00:00:16,100 --> 00:00:19,130
Stable diffusion is simply a diffusion model.
5
00:00:19,130 --> 00:00:23,630
And in the next lecture we take a closer look what stable diffusion is.
6
00:00:23,630 --> 00:00:29,330
And we use stable diffusion because it is entirely free.
7
00:00:29,360 --> 00:00:33,770
There are a lot of options how to use stable diffusion.
8
00:00:33,770 --> 00:00:35,630
There is Leonardo AI.
9
00:00:35,630 --> 00:00:37,850
That's something I really, really like.
10
00:00:37,850 --> 00:00:40,220
There are a lot of Colab notebooks.
11
00:00:40,220 --> 00:00:43,010
They are a little bit tricky to use.
12
00:00:43,010 --> 00:00:47,990
There's also comfy UI, there's automatic 1111.
13
00:00:47,990 --> 00:00:50,690
Most of them are not that simple to use.
14
00:00:50,690 --> 00:00:53,450
You need to install Python and so on.
15
00:00:53,450 --> 00:00:56,060
But we will take the shortcut.
16
00:00:56,060 --> 00:01:04,400
We will take cod because cod is what is in my mind one of the easiest ways to use stable diffusion.
17
00:01:04,400 --> 00:01:10,730
Stable diffusion is simply the model that allows us to make our pictures so stable.
18
00:01:10,730 --> 00:01:18,620
Diffusion is the diffusion model and Cod is simply the platform we will use Laura's, we will use seeds,
19
00:01:18,620 --> 00:01:24,710
we will use checkpoints and and different models to make perfect pictures that look similar.
20
00:01:24,710 --> 00:01:32,540
Like I said, if the pictures are not that perfect, we will use a technique that's called face swap.
21
00:01:32,540 --> 00:01:40,940
We will use Google Colab to swap our faces, and after that process we will have consistent pictures.
22
00:01:40,940 --> 00:01:43,640
They are nearly perfect after this.
23
00:01:43,640 --> 00:01:46,610
So that's what we do over the next lectures.
24
00:01:46,610 --> 00:01:50,600
First we take a look what exactly is stable diffusion.
25
00:01:50,600 --> 00:01:53,300
So what is a diffusion model and how it works.
26
00:01:53,300 --> 00:01:55,190
Then we take a closer look.
27
00:01:55,190 --> 00:01:58,910
So what are Laura's, what are seeds, what are checkpoints.
28
00:01:58,910 --> 00:01:59,840
And so on.
29
00:01:59,840 --> 00:02:05,540
We make our pictures and we use perfect prompt engineering for our pictures.
30
00:02:05,540 --> 00:02:09,830
We make our influencer in different environments.
31
00:02:09,830 --> 00:02:11,960
And then the fun stuff begins.
32
00:02:11,960 --> 00:02:16,850
Because after that we will use Google Colab and make some face swaps.
33
00:02:16,850 --> 00:02:19,400
So have fun over the next lectures.
34
00:02:19,790 --> 00:02:26,630
And first we take a closer look what stable diffusion exactly is and how the diffusion model works,
35
00:02:26,630 --> 00:02:30,590
because we need this for the prompt engineering later.
3312
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.