Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
0
00:00:00,200 --> 00:00:04,790
I've been using Leonardo AI for the past
few weeks as I'm trying to find the best
1
00:00:04,800 --> 00:00:09,750
alternatives to Mid Journey, and in this
video I'll be covering the benefits of
2
00:00:09,760 --> 00:00:13,000
using Leonardo AI for creating coloring
books.
3
00:00:13,040 --> 00:00:16,630
Now I've been taking the time to look
through different text to image tools,
4
00:00:16,640 --> 00:00:21,350
and Leonardo AI definitely has some
unique features that sets itself apart
5
00:00:21,360 --> 00:00:22,400
from other tools.
6
00:00:22,480 --> 00:00:24,510
And especially if you're a beginner, you
7
00:00:24,520 --> 00:00:29,550
like the fact that Leonardo AI is
actually on a website rather than being
8
00:00:29,560 --> 00:00:31,880
on. This scored where it can be quite messy
9
00:00:31,890 --> 00:00:33,440
for somebody just starting out.
10
00:00:33,450 --> 00:00:35,520
I actually really like the layout of this
11
00:00:35,530 --> 00:00:38,200
website. I think it's really easy to use and
12
00:00:38,210 --> 00:00:41,040
understand. Now, what Leonardo is best known for is
13
00:00:41,050 --> 00:00:45,840
actually creating a game assets and
concept art, but this video will
14
00:00:45,850 --> 00:00:47,840
specifically be on coloring books.
15
00:00:47,850 --> 00:00:50,250
Now getting on to explain the homepage,
16
00:00:50,450 --> 00:00:53,210
the first thing you'll see is the
featured models.
17
00:00:53,290 --> 00:00:57,650
Now this is a unique feature that allows
you to select one of the models closest
18
00:00:57,970 --> 00:01:00,010
to the style that you want to generate.
19
00:01:00,250 --> 00:01:02,160
This is very beginner friendly as you
20
00:01:02,170 --> 00:01:05,680
don't have to struggle to come up with a
particular style, but you can use one of
21
00:01:05,690 --> 00:01:08,960
the models.
So the model I like to use the most is
22
00:01:08,970 --> 00:01:12,640
Leonardo Diffusion, so you can just click
on that, then you can have a read through
23
00:01:12,650 --> 00:01:13,370
the description.
24
00:01:13,690 --> 00:01:15,280
You can also take a look at images
25
00:01:15,290 --> 00:01:18,560
created with this model and if it's
something that you like then you can just
26
00:01:18,570 --> 00:01:20,760
click generate with this model.
27
00:01:20,770 --> 00:01:22,120
But I'll take you through image
28
00:01:22,130 --> 00:01:26,690
generation a little bit later and lower
down you can see the recent creations.
29
00:01:27,090 --> 00:01:29,970
So we've got an option to see what's
trending what's.
30
00:01:30,090 --> 00:01:35,210
Newly been created and what's top so some
of the best things that people have liked
31
00:01:35,250 --> 00:01:40,010
and it's been created using Leonardo AI
and you've got an option to see all or
32
00:01:40,090 --> 00:01:43,440
game assets.
So if you are on a free plan then your
33
00:01:43,450 --> 00:01:46,800
images are public and anybody can see
them.
34
00:01:46,810 --> 00:01:51,120
So let's say people are able to find your
image in the new and they start to like
35
00:01:51,130 --> 00:01:54,240
it and it could potentially become
trending and down the line it could
36
00:01:54,250 --> 00:01:56,560
eventually become one of the top images.
37
00:01:56,570 --> 00:01:58,280
If you're thinking about selling your
38
00:01:58,290 --> 00:02:00,040
images then that's probably something.
39
00:02:00,050 --> 00:02:01,200
That you won't want to happen.
40
00:02:01,210 --> 00:02:04,760
So what I would recommend is actually
deleting your image if you're on the free
41
00:02:04,770 --> 00:02:09,560
plan or going onto one of the paid
subscriptions that will provide you with
42
00:02:09,570 --> 00:02:12,560
private mode.
But here what we can do is type in
43
00:02:12,570 --> 00:02:16,320
coloring pages just to take a look at
some of the images that users have
44
00:02:16,330 --> 00:02:18,640
created. Then you can actually take a scroll
45
00:02:18,650 --> 00:02:21,240
through and any particular image that you
do like.
46
00:02:21,250 --> 00:02:23,450
You'll be provided with a few options.
47
00:02:23,530 --> 00:02:24,840
Now again, if you are doing this, I
48
00:02:24,850 --> 00:02:27,920
wouldn't recommend taking anybody's image
and copying it.
49
00:02:27,930 --> 00:02:29,810
We want to create unique work.
50
00:02:30,050 --> 00:02:31,640
That will make our work stand out on
51
00:02:31,650 --> 00:02:36,440
Amazon at KDP, but there's nothing wrong
being inspired by somebody's work and
52
00:02:36,450 --> 00:02:38,400
looking at ideas from their prompt.
53
00:02:38,410 --> 00:02:40,280
You can also see what model the image was
54
00:02:40,290 --> 00:02:44,360
generated with so that you get a better
idea of the best models to use for
55
00:02:44,370 --> 00:02:46,880
coloring books.
So going through the left hand grid,
56
00:02:46,890 --> 00:02:48,240
there's quite a few options here.
57
00:02:48,250 --> 00:02:50,520
Just under home we've got community feed,
58
00:02:50,530 --> 00:02:54,600
and again, you can take a look at all the
designs created by the community and you
59
00:02:54,610 --> 00:02:57,520
can search for any particular thing
you're looking for.
60
00:02:57,530 --> 00:03:01,240
Next, we've got the personal feed so you
can take a look at the designs that you
61
00:03:01,250 --> 00:03:02,770
have already created.
62
00:03:03,330 --> 00:03:04,960
After this we've got training and
63
00:03:04,970 --> 00:03:07,640
datasets. And another thing which makes Leonardo AI
64
00:03:07,650 --> 00:03:12,600
really unique is the train and create a
particular model and this will allow you
65
00:03:12,610 --> 00:03:17,360
to easily generate images based on the
style that you're looking for and I'll
66
00:03:17,370 --> 00:03:18,680
show you that a little bit later.
67
00:03:18,690 --> 00:03:21,450
Then we've also got fine-tuned models, so
68
00:03:21,530 --> 00:03:24,720
again you can use any of these to base
your image on.
69
00:03:24,730 --> 00:03:30,040
You also have community models, so these
were created by users in the.
70
00:03:30,050 --> 00:03:33,680
Unity and you can take a look through
them and use any particular model of your
71
00:03:33,690 --> 00:03:36,760
interest. Then the next section is the user tools
72
00:03:36,770 --> 00:03:41,000
and the first is the AI image generation,
which is where you're going to be
73
00:03:41,010 --> 00:03:43,040
creating all of your images.
74
00:03:43,050 --> 00:03:45,000
Then we've got the AI canvas and this
75
00:03:45,010 --> 00:03:48,010
will allow you to actually edit your AI
image.
76
00:03:48,130 --> 00:03:51,720
Then we've got texture generation, so
that's mostly for 3D models, so we won't
77
00:03:51,730 --> 00:03:53,370
be focusing on that on our video.
78
00:03:53,450 --> 00:03:56,280
Then you've got settings and FAQ and
79
00:03:56,290 --> 00:03:58,080
help. So the first part I want to go into a
80
00:03:58,090 --> 00:04:02,410
little bit more is the training and
datasets now this will allow you to
81
00:04:02,610 --> 00:04:06,450
easily generate images based on the style
that you're looking for.
82
00:04:06,570 --> 00:04:10,520
So in this instance what I did was go
into creative Fabric and download some
83
00:04:10,530 --> 00:04:15,000
mandala coloring pages just to test it
out and to be able to have my own model
84
00:04:15,010 --> 00:04:17,410
where I could create mandala coloring
pages.
85
00:04:17,529 --> 00:04:22,520
So to test this out all you need to do is
click on new dataset and then you need to
86
00:04:22,530 --> 00:04:25,440
enter a name and description for the
dataset.
87
00:04:25,450 --> 00:04:27,050
Then you click create dataset.
88
00:04:27,330 --> 00:04:29,970
Next you can start to upload your images.
89
00:04:30,050 --> 00:04:33,120
So I uploaded about 40 man dollar
coloring pages.
90
00:04:33,130 --> 00:04:37,120
I think the more the better and as soon
as all those have been uploaded then
91
00:04:37,130 --> 00:04:41,000
you're going to have the option to click
train model and you will see some of
92
00:04:41,010 --> 00:04:42,240
these settings come up.
93
00:04:42,250 --> 00:04:43,840
And most of these are just down to your
94
00:04:43,850 --> 00:04:46,040
personal preference.
So you can choose your training
95
00:04:46,050 --> 00:04:50,370
resolution category.
The base model it usually recommends
96
00:04:50,570 --> 00:04:53,640
stable diffusion 1 5 and then an instance
prompt.
97
00:04:53,650 --> 00:04:56,280
Here you're just describing your image in
two to three words.
98
00:04:56,290 --> 00:04:59,400
Actually use this up on the free plan as
I think you're only able to create one to
99
00:04:59,410 --> 00:05:00,120
two models.
100
00:05:00,130 --> 00:05:02,880
A month, but it will reset in 22 days and
101
00:05:02,890 --> 00:05:07,400
once you click start training, Leonardo
will go away and train the model and as
102
00:05:07,410 --> 00:05:11,720
soon as it's completed then you'll get an
email notifying you that your model is
103
00:05:11,730 --> 00:05:14,400
ready. So to start generating your images, make
104
00:05:14,410 --> 00:05:19,600
sure you go to the AI image generation
and once that loads you'll be taken to
105
00:05:19,610 --> 00:05:24,200
this screen and if you look in the far
left hand corner you're actually going to
106
00:05:24,210 --> 00:05:27,040
be able to see how many tokens you have
left.
107
00:05:27,050 --> 00:05:30,040
So the way Leonardo AI works is that you.
108
00:05:30,050 --> 00:05:32,960
Have a certain limit of tokens, and each
109
00:05:32,970 --> 00:05:37,360
time you generate an image or upscale
this will actually use up a certain
110
00:05:37,370 --> 00:05:38,690
number of tokens.
111
00:05:38,850 --> 00:05:40,320
So if you're in the free version, you
112
00:05:40,330 --> 00:05:46,400
only have one fifty tokens to use a day
and then it actually resets at the end of
113
00:05:46,410 --> 00:05:46,960
each day.
114
00:05:46,970 --> 00:05:49,000
You really need to be aware of how many
115
00:05:49,010 --> 00:05:53,240
tokens you're using each time you
generate an image, and on the free
116
00:05:53,250 --> 00:05:59,080
version you're actually able to generate
a maximum of four images and the next
117
00:05:59,090 --> 00:06:00,040
part is prompt.
118
00:06:00,050 --> 00:06:02,920
To magic and this actually allows you to
119
00:06:02,930 --> 00:06:04,890
have better prompt adherence.
120
00:06:04,970 --> 00:06:08,600
So turning this on will allow the image
121
00:06:08,610 --> 00:06:10,410
to be much closer to your prompt.
122
00:06:10,570 --> 00:06:12,400
Now I didn't have this, but this will
123
00:06:12,410 --> 00:06:16,440
allow you to adjust the strength, but
it's something you might want to test out
124
00:06:16,450 --> 00:06:18,640
if you're not getting the images that you
want.
125
00:06:18,650 --> 00:06:23,480
After this we have the image dimension so
you can choose the best dimensions for
126
00:06:23,490 --> 00:06:29,600
your image.
I went with 768 by 1024 for coloring
127
00:06:29,610 --> 00:06:32,520
page. Images and we've got the guidance scale
128
00:06:32,530 --> 00:06:37,200
after this and this allows you to adjust
how strongly your prompt is weighted, but
129
00:06:37,210 --> 00:06:42,600
I usually just leave this on 7 after we
have control net and this allows you to
130
00:06:42,610 --> 00:06:45,130
influence your generations with input
images.
131
00:06:45,290 --> 00:06:50,850
However, this currently works only with
models derived from a stable diffusion 1
132
00:06:51,010 --> 00:06:53,880
5. And next we have tiling and you can turn
133
00:06:53,890 --> 00:06:56,970
this on if you want repeated patterns and
backgrounds.
134
00:06:57,210 --> 00:07:00,040
Finally you've got image to image and
that just allows you to use.
135
00:07:00,050 --> 00:07:03,240
An image as a reference, but you still
have to type in the prompt and then
136
00:07:03,250 --> 00:07:07,960
you've also got image prompt and that
allows you to use an existing image to
137
00:07:07,970 --> 00:07:10,560
describe how the generations should look.
138
00:07:10,570 --> 00:07:12,480
With either one you just need to upload
139
00:07:12,490 --> 00:07:14,920
or you can drag and drop the image into
here.
140
00:07:14,930 --> 00:07:16,000
So at the top.
141
00:07:16,010 --> 00:07:17,160
This is where you're going to be able to
142
00:07:17,170 --> 00:07:21,680
type in your prompt, and then just
underneath you can choose the model that
143
00:07:21,690 --> 00:07:22,970
you want to use.
144
00:07:23,410 --> 00:07:25,920
So if you want to use the custom model
145
00:07:25,930 --> 00:07:30,080
that you've already made, you just can
select that at the bottom and then you.
146
00:07:30,090 --> 00:07:34,200
Click View and then you can generate with
that model and then I'll actually save it
147
00:07:34,210 --> 00:07:35,960
there. So if you ever want to generate a
148
00:07:35,970 --> 00:07:39,640
specific image with that model then it's
already saved there at the top so you can
149
00:07:39,650 --> 00:07:41,290
see mandala coloring pages.
150
00:07:41,620 --> 00:07:43,210
Then you can choose if you want Leonardo
151
00:07:43,220 --> 00:07:44,170
style or none.
152
00:07:44,180 --> 00:07:46,330
I always go with Leonardo style and if
153
00:07:46,340 --> 00:07:49,330
you want to add a negative prompt you
just turn that on.
154
00:07:49,340 --> 00:07:53,340
This allows you to tell Leonardo what you
don't want to see within the image.
155
00:07:53,660 --> 00:07:56,850
Then you've got generate on the right
hand side and as soon as you click that
156
00:07:56,860 --> 00:07:58,380
it will generate your image.
157
00:07:58,540 --> 00:08:01,690
Now depending on the number of images
158
00:08:01,700 --> 00:08:06,090
that you use and also the model, that
will depend on how many tokens you use
159
00:08:06,100 --> 00:08:08,530
up. So in this instance we're using four
160
00:08:08,540 --> 00:08:11,580
images and the Leonardo diffusion model.
161
00:08:11,620 --> 00:08:13,740
And that's going to use up five tokens.
162
00:08:13,780 --> 00:08:17,890
However, if I decrease that down to two
images, you can see that goes down to
163
00:08:17,900 --> 00:08:20,770
three tokens.
So especially if you're on the free plan,
164
00:08:20,780 --> 00:08:25,130
just be wise with how many tokens you use
because you only have a certain limit
165
00:08:25,140 --> 00:08:27,650
each day.
Now the awesome feature of Leonardo is
166
00:08:27,660 --> 00:08:31,570
this prompt generation, so if you are
struggling to come up with prompts, then
167
00:08:31,580 --> 00:08:34,539
you can choose the number of prompts that
you want to generate.
168
00:08:34,620 --> 00:08:39,490
You've got 2-4-6-8 and so on and you have
a sizable amount that you can use.
169
00:08:39,500 --> 00:08:41,610
You can see here I've still got 900 and.
170
00:08:41,620 --> 00:08:43,049
78 prompts remaining.
171
00:08:43,059 --> 00:08:48,970
So let's say I type in a lion coloring
page and I choose six prompts to generate
172
00:08:48,980 --> 00:08:53,730
and I just click ID 8 here and then it's
going to come up with six different ideas
173
00:08:53,740 --> 00:08:54,330
for me.
174
00:08:54,340 --> 00:08:56,410
Again, this isn't always perfect, You can
175
00:08:56,420 --> 00:09:00,170
see only one of them actually include the
word coloring page, but you can get a lot
176
00:09:00,180 --> 00:09:03,610
of ideas from this in order to add to
your prompt.
177
00:09:03,620 --> 00:09:09,370
Then for any particular prompt that you
do like, just copy it and then go back to
178
00:09:09,380 --> 00:09:11,610
image generation, paste it in the.
179
00:09:11,620 --> 00:09:13,810
Any type of prompt, make sure you choose
180
00:09:13,820 --> 00:09:17,890
the model that you want and then you can
click generate and it will come up with
181
00:09:17,900 --> 00:09:19,770
the amount of images that you've chosen.
182
00:09:19,780 --> 00:09:21,570
So let's quickly go to my personal feed
183
00:09:21,580 --> 00:09:25,060
and just take a look at some of the
coloring pages that we've generated.
184
00:09:25,100 --> 00:09:29,090
So at the bottom we can see some mandala
pages and we've got some owl coloring
185
00:09:29,100 --> 00:09:31,620
pages that look really good.
186
00:09:31,660 --> 00:09:33,770
We've got a golden retriever coloring
187
00:09:33,780 --> 00:09:35,300
page. We've got some birds.
188
00:09:35,770 --> 00:09:41,920
We've got a lion portrait and then at the
top we've got some more Mandala pages.
189
00:09:41,930 --> 00:09:44,840
So there are definitely some quality
coloring pages here.
190
00:09:44,850 --> 00:09:48,680
However, there's others really aren't any
great at all and that just really depends
191
00:09:48,690 --> 00:09:50,040
on the model that you use.
192
00:09:50,050 --> 00:09:52,440
Now going back to the AI generation tool,
193
00:09:52,450 --> 00:09:56,880
we can take a look at these, our coloring
pages sitting on a branch.
194
00:09:56,890 --> 00:09:57,840
So that was the prompt.
195
00:09:57,850 --> 00:09:59,720
And then we can see that the model that
196
00:09:59,730 --> 00:10:03,720
we've used is Leonardo diffusion and
that's Leonardo style and there's four
197
00:10:03,730 --> 00:10:05,760
images here.
Now these came out.
198
00:10:05,770 --> 00:10:06,400
Quite well.
199
00:10:06,410 --> 00:10:09,680
However the 1st and the third image have
200
00:10:09,690 --> 00:10:10,640
a black background.
201
00:10:10,650 --> 00:10:13,690
So what I did in the next image prompt
202
00:10:13,730 --> 00:10:19,120
was to put in a negative prompt for black
background and that got rid of the black
203
00:10:19,130 --> 00:10:22,120
background in these images as you can see
here.
204
00:10:22,130 --> 00:10:27,520
Then I tried it with the golden retriever
sitting and it came up with some really
205
00:10:27,530 --> 00:10:32,480
good images again with Leonardo diffusion
and taking a look further up you can see
206
00:10:32,490 --> 00:10:35,760
we've got some birds on a branch coloring
page and these.
207
00:10:35,770 --> 00:10:39,800
This came out quite good apart from the
fourth one, but sometimes AI does
208
00:10:39,810 --> 00:10:41,530
generate some strange images.
209
00:10:41,570 --> 00:10:44,920
Going further up we've got a lion kind of
210
00:10:44,930 --> 00:10:49,730
portrait coloring page and this came out
quite blue for some reasons.
211
00:10:50,050 --> 00:10:54,850
I used the image prompt generation tool
to provide me with this prompt.
212
00:10:55,010 --> 00:11:01,610
So then in the next part what I decided
to do was to include white background
213
00:11:01,770 --> 00:11:05,800
within the main prompt and this came out
much better.
214
00:11:05,810 --> 00:11:08,960
However, I noticed that some of the
image, especially two and four, were
215
00:11:08,970 --> 00:11:10,210
still quite blue.
216
00:11:10,250 --> 00:11:13,800
So again, in the next prompt, what I did
217
00:11:13,810 --> 00:11:19,760
was include blue background within the
negative prompt and these all came out
218
00:11:19,770 --> 00:11:21,960
really good.
And that's sometimes what you're going to
219
00:11:21,970 --> 00:11:23,930
have to do with prompts generating.
220
00:11:24,050 --> 00:11:25,200
It's not going to come up with the
221
00:11:25,210 --> 00:11:26,560
perfect image every time.
222
00:11:26,570 --> 00:11:28,520
You might have to just add a few things
223
00:11:28,530 --> 00:11:30,600
and mess about with a few things here and
there.
224
00:11:30,610 --> 00:11:35,200
Now, in terms of the front cover for the
coloring book, I took out the words
225
00:11:35,210 --> 00:11:38,280
coloring. Page as I didn't want it to be black and
226
00:11:38,290 --> 00:11:41,880
white and actually added in the words
colorful just so I could get an
227
00:11:41,890 --> 00:11:44,000
interesting image as the front cover.
228
00:11:44,010 --> 00:11:45,800
And it was actually number three that I
229
00:11:45,810 --> 00:11:48,480
particularly liked the most and will use
as a front cover.
230
00:11:48,490 --> 00:11:53,960
Now these Mandala coloring pages were
actually created using my own model, the
231
00:11:53,970 --> 00:11:59,720
Mandala Coloring page, and you can see
the an example of the coloring pages that
232
00:11:59,730 --> 00:12:02,000
I uploaded and how they turned out.
233
00:12:02,010 --> 00:12:04,320
Now I wasn't too impressed with these,
234
00:12:04,330 --> 00:12:05,760
but I'll still try and train.
235
00:12:05,770 --> 00:12:08,080
Another model and see how that turns out.
236
00:12:08,090 --> 00:12:12,960
So don't be afraid to test Leonardo out,
but I'll definitely stick with the model
237
00:12:12,970 --> 00:12:16,720
Leonardo Diffusion as I tested out the
others and they just weren't good enough
238
00:12:16,730 --> 00:12:18,810
specifically for coloring pages.
239
00:12:18,890 --> 00:12:20,480
And to complete your coloring book, I
240
00:12:20,490 --> 00:12:22,810
would at least have 20 images.
241
00:12:22,930 --> 00:12:25,160
Make sure the images that you do use are
242
00:12:25,170 --> 00:12:30,080
upscaled, so there's actually quite a few
options within Leonardo to upscale an
243
00:12:30,090 --> 00:12:34,880
image, and if you go into the image you
can see quite a few different options to
244
00:12:34,890 --> 00:12:37,720
choose from.
So at the top you've got creative upscale
245
00:12:37,730 --> 00:12:42,960
and it says this can improve images
during the upscaling process, but it will
246
00:12:42,970 --> 00:12:44,650
cost you 5 tokens.
247
00:12:44,930 --> 00:12:46,680
Then to the left hand side you've got HD
248
00:12:46,690 --> 00:12:49,760
Smooth Upscale.
So this is an upscaler that works well
249
00:12:49,770 --> 00:12:54,600
with a focused subject but can end up
smoothing out fine detail and again
250
00:12:54,610 --> 00:12:56,010
costing you 5 tokens.
251
00:12:56,130 --> 00:12:59,010
In the middle we've got the upscale image
252
00:12:59,050 --> 00:13:04,120
alternate and you can use this if you
find a creative upscale that is resulting
253
00:13:04,130 --> 00:13:05,890
in a loss of detail.
254
00:13:06,170 --> 00:13:08,410
And again, this is costing you 5 tokens.
255
00:13:08,530 --> 00:13:12,760
On the right hand side we've got the use
for image to image, which you won't need
256
00:13:12,770 --> 00:13:15,960
right now.
Then at the bottom we've got HD Crisp
257
00:13:15,970 --> 00:13:20,280
upscale and this is a great balanced
upscaler which retains a good amount of
258
00:13:20,290 --> 00:13:22,640
detail with crispiness to the image.
259
00:13:22,650 --> 00:13:24,890
And again that's five tokens.
260
00:13:25,130 --> 00:13:27,880
Then we also have an option to remove the
background.
261
00:13:27,890 --> 00:13:33,440
This will cost you 2 tokens and to the
left we've got in unzoom image which will
262
00:13:33,450 --> 00:13:35,760
cost you 5 tokens and to.
263
00:13:35,770 --> 00:13:38,360
Right, we've got edit in canvas, which
264
00:13:38,370 --> 00:13:39,810
I'll show you a little bit later.
265
00:13:40,050 --> 00:13:42,200
So the two biggest issues that I had with
266
00:13:42,210 --> 00:13:48,320
upscaling in Leonardo is one that it does
cost quite a few tokens to upscale only
267
00:13:48,330 --> 00:13:51,160
one image.
And think about it with a coloring book.
268
00:13:51,170 --> 00:13:56,000
We are using a minimum of 20 images, so
20 times 5.
269
00:13:56,010 --> 00:14:00,520
You're at least using a minimum of 100
tokens and this is just not going to be
270
00:14:00,530 --> 00:14:04,440
enough, especially if you are using the
free plan because you also need to
271
00:14:04,450 --> 00:14:05,760
generate the images in.
272
00:14:05,770 --> 00:14:08,240
1st place and the second point is even
273
00:14:08,250 --> 00:14:12,360
when you do upscale these images in
Leonardo, they will still need to be
274
00:14:12,370 --> 00:14:16,840
upscaled within an external website as
coloring pages do have a lot of intricate
275
00:14:16,850 --> 00:14:23,440
details and I found if you upscale it in
Leonardo and using an external tool then
276
00:14:23,450 --> 00:14:28,080
the files just become way too big south
that's why I would avoid upscaling in
277
00:14:28,090 --> 00:14:29,890
Leonardo for coloring books.
278
00:14:30,010 --> 00:14:31,800
The actual website that I would recommend
279
00:14:31,810 --> 00:14:35,760
that you upscale your images in is big
jpg.
280
00:14:35,770 --> 00:14:41,440
Dot com It has very low cost and
affordable plans which you can see here.
281
00:14:41,450 --> 00:14:45,760
You can also try it and just test it out
on the free plan so that there's no
282
00:14:45,770 --> 00:14:48,720
commitment. But then as soon as you're happy with it,
283
00:14:48,730 --> 00:14:53,720
you can actually batch upload your images
and you just click select and choose all
284
00:14:53,730 --> 00:14:55,080
the images you want to upload.
285
00:14:55,090 --> 00:14:56,290
Then just click start.
286
00:14:56,450 --> 00:15:01,840
Make sure you choose at least four times,
upscaling as 8 and 16 times the files
287
00:15:01,850 --> 00:15:05,770
become way too big and then you can just
go down to enlarge.
288
00:15:05,930 --> 00:15:08,960
All images below and once that's finished
then click download.
289
00:15:08,970 --> 00:15:12,160
Now I quickly wanted to show you the
difference between both images that have
290
00:15:12,170 --> 00:15:14,920
been upscaled.
On the left hand side we've got the image
291
00:15:14,930 --> 00:15:20,240
that's been upscaled in Leonardo AI using
the HD Crisp and on the right hand side
292
00:15:20,250 --> 00:15:23,450
we've got the image has been upscaled
using bigjpg.com
293
00:15:24,650 --> 00:15:28,640
Now we can see when we zoom in on the
left hand side that it begins to lose
294
00:15:28,650 --> 00:15:33,960
some detail whereas on the right hand
side you can see how crisp and detailed
295
00:15:33,970 --> 00:15:34,760
the image stays.
296
00:15:34,770 --> 00:15:35,770
So that's the important.
297
00:15:36,050 --> 00:15:41,850
ofusingbigjpg.com as we want to give the
via the best coloring experience.
298
00:15:42,250 --> 00:15:47,200
Another awesome tool Leonardo AI has is
the AI Canvas editor and once we enter it
299
00:15:47,210 --> 00:15:50,480
we're provided with a few options on the
left and right hand side.
300
00:15:50,490 --> 00:15:55,640
On the left hand side we have firstly the
pan tool and this will allow you to move
301
00:15:55,650 --> 00:15:56,890
the image around.
302
00:15:57,090 --> 00:15:59,850
Then underneath we've got the snap tool.
303
00:16:00,010 --> 00:16:03,720
Then underneath that we've got draw a
mask, we've got arrays, we've got upload
304
00:16:03,730 --> 00:16:05,800
image and then underneath that we've got.
305
00:16:05,810 --> 00:16:08,610
Download and undo and redo.
306
00:16:08,690 --> 00:16:11,600
In the middle we've got where we can
enter our prompt and then to the right
307
00:16:11,610 --> 00:16:14,000
hand side we've got image dimensions and
a few other options.
308
00:16:14,010 --> 00:16:19,240
To start off we can upload our image and
use the pan tool to move the image around
309
00:16:19,250 --> 00:16:24,240
to where we want and then we can also use
the zoom up at the top to help us see the
310
00:16:24,250 --> 00:16:25,410
image a lot better.
311
00:16:25,570 --> 00:16:27,400
Next what we need to do is the image is a
312
00:16:27,410 --> 00:16:28,640
lot larger than the box.
313
00:16:28,650 --> 00:16:30,560
So what we can do is click on the snap
314
00:16:30,570 --> 00:16:35,760
tool and then change the image dimensions
to seven sixty eight by ten.
315
00:16:35,770 --> 00:16:39,410
24 as that's the coloring page size and
then just rearrange it so everything
316
00:16:39,450 --> 00:16:42,120
aligns properly.
Now one thing I've noticed on this image
317
00:16:42,130 --> 00:16:44,440
is that the branch doesn't go all the way
to the edge.
318
00:16:44,450 --> 00:16:49,280
So what I'm going to do is use the draw
mask tool and you can increase the size
319
00:16:49,290 --> 00:16:53,480
of it and I'm going to go to the area and
then I'm going to draw over that specific
320
00:16:53,490 --> 00:16:55,530
area that I want to make a change.
321
00:16:55,610 --> 00:16:57,410
Then we can just go down to the bottom
322
00:16:57,450 --> 00:17:00,040
where it has the prompt and then we type
in what we want.
323
00:17:00,050 --> 00:17:04,760
So then I just put in expand branch, then
I'll click generate, give it a moment and
324
00:17:04,770 --> 00:17:05,800
once it's done.
325
00:17:05,810 --> 00:17:07,599
Provide you with four different options.
326
00:17:07,609 --> 00:17:13,319
Now this does cost 5 tokens and then you
can scroll across just to see which one
327
00:17:13,329 --> 00:17:15,560
that you like the most and I was happy
with three.
328
00:17:15,569 --> 00:17:18,130
It looked most similar to the image and
then just click accept.
329
00:17:18,470 --> 00:17:22,579
And once you're happy with everything,
then you can download your image.
330
00:17:22,589 --> 00:17:25,660
So I'm going to take a look at the terms
and conditions so you understand what you
331
00:17:25,670 --> 00:17:30,460
are able to do with Leonardo AI and it's
fantastic that it does state that you own
332
00:17:30,470 --> 00:17:34,140
all the content you generate using the
platform and that yes, you can use the
333
00:17:34,150 --> 00:17:37,110
images generated by the platform for
commercial purposes.
334
00:17:37,310 --> 00:17:40,620
And this applies to images created by
free users too.
335
00:17:40,630 --> 00:17:44,870
So this is a great option for those who
aren't able to commit to a subscription
336
00:17:45,310 --> 00:17:48,030
right now, but taking a look at their
subscriptions.
337
00:17:48,470 --> 00:17:53,580
You can see if you're paying monthly
you've got a 12$ a month with some great
338
00:17:53,590 --> 00:17:56,820
options. You've got a 30$ a month with even more
339
00:17:56,830 --> 00:18:00,940
great options and then the 60$ a month
plan for those who need it.
340
00:18:00,950 --> 00:18:04,820
Now the main benefits of being on a
subscription is that you're able to do a
341
00:18:04,830 --> 00:18:09,780
lot more pending jobs, that you also have
private generations as well so nobody can
342
00:18:09,790 --> 00:18:10,950
see what you're designing.
343
00:18:11,030 --> 00:18:12,380
Then you've also got priority
344
00:18:12,390 --> 00:18:16,020
infrastructure as well, and then with the
higher plans you've got relaxed
345
00:18:16,030 --> 00:18:18,580
generation queue also, so take a.
346
00:18:18,590 --> 00:18:20,340
And consider what plan you want to go on.
347
00:18:20,350 --> 00:18:25,260
Or if you want to subscribe to the year
of to a yearly plan then you get 20 %
348
00:18:25,270 --> 00:18:26,740
off. So you can just take a look through
349
00:18:26,750 --> 00:18:29,980
those. And if Leonardo AI is something that you
350
00:18:29,990 --> 00:18:33,700
want to sign up to, then you just need to
go to their website and then put your
351
00:18:33,710 --> 00:18:36,710
email. In then you get access to their Discord
352
00:18:37,070 --> 00:18:39,870
and then make sure you go down to a
priority early access.
353
00:18:39,950 --> 00:18:42,070
Then you just need to fill out a form.
354
00:18:42,190 --> 00:18:44,540
Once you've submitted the form, then in
355
00:18:44,550 --> 00:18:48,460
the next few days they'll get back to you
and give you access to Leonardo.
356
00:18:48,470 --> 00:18:53,420
Ai so to finish off this I just wanted to
show you how the coloring book turned out
357
00:18:53,430 --> 00:18:56,630
using a book bolt so you can see that's
the cover I used.
358
00:18:56,670 --> 00:19:01,150
Then just a simple title and subtitle and
I've got the images on the back cover.
359
00:19:01,550 --> 00:19:07,140
Then on the title page I've just got the
coloring page image with again the title
360
00:19:07,150 --> 00:19:10,660
subtitle. Then always leave the next page blank and
361
00:19:10,670 --> 00:19:14,660
then have your next coloring image on the
next page after that then leave it blank
362
00:19:14,670 --> 00:19:16,580
again and then have your coloring image.
363
00:19:16,590 --> 00:19:18,460
So it's just following that process and.
364
00:19:18,470 --> 00:19:22,620
Make sure you have at least 20 images in
your coloring book.
365
00:19:22,630 --> 00:19:26,500
Then once you're hay with everything,
just go to download and make sure you
366
00:19:26,510 --> 00:19:29,280
download it as a CMYK that's best 4
rinted.
33483
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.