Transcripts
1. Introduction: Generated all these
photos of myself with a Cava app that nobody is talking about.
I mean, look at them. They look so real
that my friends and even my own mother had a hard time figuring out
what was going on here. So, in this video, I'll show you everything
I know about this app. I'll tell you what it
is, how to use it, and give you my
best tips so that you too can get the
best photos out of it.
2. Step 1: Training Your AI Model: The very first thing
you will need to do is to come to Canva and
locate the apps button. Okay, so click on
Apps and then search for headshots,
headshots, like this. And the logo looks like this. It's a camera icon,
so click on it, and you should get to this menu right here.
So there are two tabs. The first one is generate and
the second one is to manage your different headshots that have been taken with the so, you need to start by
training a new model. So we can see here the app prompts us to
upload some photos. So upload at least ten photos where only your face is visible. Uploading more photos will
provide better results. Then if we look at the
second field right here, the model name is simply a field for you to
name your mother, to give that mother that
you're going to train a me, I have decided to
train the model on my photos to be
fully transparent. I already have a
model trained here, and I want to train
a new one because since I trained my first model, I talked to my
friend at headshot, and he gave me some tips on
how to optimize the model, which kind of photos to upload so that you get the
best possible results. So I want to try this new way and also show you the
steps, obviously. So what else should you know
about training a model? Well, first of all,
it's not free. Like I was telling you before, it will cost you
credits 90 credits to train a new model based
on a new set of photos. Included in this
AI model training, you will get 40
headshots generated automatically as soon as the model stops,
finishes the training. Okay? So once it's trained, it will generate 40
headshots for you, you will receive an
email letting you know, Hey, your model is ready,
discover your headshot. So we'll go through this. The training takes
about an hour. Okay, so it takes a
bit of obviously, these credits need to be
purchased at some point. So how do you get the credits? You need to click on that
little credit right here, Link. So let's open that.
There you go. You will get to this page
right here where you can choose between three
different packs of credits. You have the basic
pack that cost $19 that will give
you 100 credit. Remember, training
the model costs 90 credits that
includes 40 heads. So this means that for $19, you can actually get a
model trained and receive 50 headshots because
every additional headshot will cost you one credit. So when you prompt, you'll
get four different images, kind of like magic media, and you can choose
any one of these. That generation of
four new images will cost you one credit. So since it's 90 credit just
for training the model, you still have ten credits to generate photos with if
you buy the smaller pack. Now, I told you I
partnered with headshots, and I managed to get you 10% off if you use the
coupon Team Rondi. So I'll show you how to
apply this in a second. Let me just explain
the other packages. So you have a $29
package for 200 crates, and a $39 package for 300. Credits. So it really
depends on your needs. If you're not sure, maybe
start with a small package, you can still buy more
credits after all, after you train your first model and you see that it's
good, that you like it, that the quality of the photos you get is consistently good, and then you can very
easily buy more credits. Continue using the 10% discount with our coupon and just
continue using the app. If you don't like it, all you would have lost is about $20. But I do think you
will like it because it's pretty so me here, I'm going to buy this $29
pack for 200 credits, get the 200 credits. So it costs you $29. But if you use our
promo code, Team Runde, Team Rundi all caps, everything attached,
you should get 10% off. Here, that's $2.90, and it
will all come down to $26. So the 10% off will be
working for everyone. Just make sure to use Team
Runde when you check out. Let me come back to
Canva and show you some tips that were
given to me by headshots on what kind of photos you should use
to train the model. So this is going to be
the second model I train, and I can show you a side by side comparison after if you want, if you're
interested into that. But first, some tips. You should only upload
photos that feature you. You should choose a set of
photos with a variety of backgrounds that will
help the AI learn better. Some recent photos, so nothing that is more than
2-years-old, maybe, include some photos that are taken in
different locations, different settings that
should also help the AI. Next, avoid exaggerated or
silly face expressions. And this is really why I decided
to train a second model. Though the first
one, I'm very happy. But I trained it on
these kinds of photos. You see the photos I usually
use on my YouTube snails, and these are
exaggerated silly faces or exaggerated
expressions on my face. So I noticed that the final output sometimes,
not all the time. And most of the time
when I generate a new photo out of the
four images generated, one is usually better
than the other ones, but the fail ones are, I believe, because I uploaded
a lot of these silly faces. So this time, I'm
going to be using much more neutral faces of mine. More variety of
background settings to help the AI really figure it out a bit
better. All right. So good lighting for clear and accurate
results showing different head
positions and angles. The first set that I used, the camera was
fixed on a tripod, so I always have kind of
the same perspective. So I'm going to
correct that with the second model I capture you looking directly at the
camera with a mix of utual expressions
and subtle smiles. So, great. Now, knowing this, I'm going to show you the sample of photos that I'm
going to be using. This should give you a good idea of the batch of images you need to gather in order to train that model the best way
possible. There you go, guys. I'm going to quickly go through my different photos. You see? So here, I have photos
without a background, like kind of neutral faces. A lot of them are
from the same batch, but then I switch it up, okay? So photos of me without the glasses, different
backgrounds, kind of different pauses, long hair, short hair,
stuff like that. This one, okay, another one. There is another
person in the photo. You might recognize
her, but let's hope the AI doesn't
get confused. Another one here with
different lighting, different hair so altogether, this is a set of 30 photos. Head Shot recommends that
you upload ten to 20 photos, but they also said that
the more the better, the more accurate
the model would be, might take slightly
longer to train. I'm not sure about that.
Also, the quality, the resolution is to be
higher than 512 pixel. To me, I try to provide the
highest possible resolution. And then last but not least, it is better to upload
your photos in PNG format. If you have the choice between
PNG and JPEG, go for PNG. But also they said that
any format just giving you the best guidelines in order to really optimize the
training of that model. Alright. So once you have
your set of photos selected, come back to headshots, and you are simply going
to start uploading. I'm going to give
this model my name. I'm just simply going
to call it Ronnie. So let's upload these photos. I'm going to click on
the Choose File button right here. This is my set. You see my entire set, so I can go Control
A or Command A to select them all open, and this will start loading
the photos right here. I have an error message, so I cannot upload all of these
photos right here because each of these images is
larger than five megabyte. Okay? So I learned
something useful here. Your photos that you upload
to headshots need to be maximum 5 megabytes large
in terms of five size. So what I need to do is to
resize all of these photos, and I'm going to show you a
tool that I used to do so. You see all of these
photos right here. I have very simply
a tool is free. It's called Shrink Media. You can resize up to ten
photos without logging in. Me, I had slightly more. So I created a free
account and haven't paid anything I'm going to show you with the last
photo remaining. So that would be
this one right here. You see the other ones are
smaller than five megabyte. So first, you need
to upload your photo into shrink Media,
this one right here. Next, it will kind of set
some random parameters, some random settings
for you, photo quality. And you see the
five size 391 kilo. Still have a lot of margin here. So I would like to push this
photo quality much higher. I'm going to go 100% in
quality and see 75% in size, see if that makes sense. No, still a little bit over. So I'm going to
slightly reduce this until I hit my sweet
spot, you know, because as Hathot
mentioned to us, the larger the higher
resolution, the better. So this 90% photo quality, 75% photo dimensioned seems to be giving me
some good results. Maybe I can push this
slightly more. Let's see. Compress. So by reducing
the quality by only 10%. That is fine. So once
you're okay with this, you can just click Download. Your photo is going
to be downloaded, and now I can upload these
two headshots one more time. So I actually
downloaded them here. So all of these will
be the new ones. You see, they are all
under 5 megabytes. So I should be able to
upload these. There you go. So now I have all of
my photos right here. I should have 30, right? I have named my model. So I can now train that model. And the way you train it is by clicking that purple
button right here. It will cost you 90 credits, but you will get 40 headshots. Alright? So let's click here. Let's start generating
these headshots and training the model. It will take a while, so I'm
going to cut the video and come back to it once
I receive that email.
3. Step 2: First Look at Your Headshots: Alright, guys, I got the email
and my model is now ready. Now, I haven't yet
discovered my 43 headshots. So I'm going to get back
to Canva right here, and I got this message. Your headshots are ready. To view them, go to manage. Okay? So let's click on Manage. I need to select the model. So remember, it's
this one right here, Ronnie, then I can
click on next. And if I head over to
headshots right here, I'm going to be able to see
my different headshots. Okay, so they are loading. Let's wait a second so we can discover them
all. All right. Interesting. So let's click
on the first one right here, make it full screen. So we can see what
really is going on here. This is pretty accurate. I would say, What do you think? As I'm saying, I'm discovering these with you
together with you. And see some of the similar outfits as per the first model
I have trained. I like this one. This
one is really cool. Different ones right
here, different outfits. Some of them, of course, don't look quite like me, but out of these 40 headshots, I can probably rescue
a good 15 to 20. All right. So I'm
not going to go over each and every one of
them individually. I just want to show you some of the good ones
and then some of the bad ones without giving you an exhaustive
tour of everything. This one doesn't look that much like me in
terms of the body. The face could be me, but that body is not my
body. What else do we have? We have this one right here. I think this one
looks pretty close. I would say head wise. Yeah, the face pretty good. I like that it gave me
different kinds of background. So this one could work.
Not my favorite one. This one is better
in my opinion. Let's go over some of the
classy ones right here. So here I have more hair. Let's have one with
more hair I like that this mother gave me
not so much like maybe like me if I had more like that this model gave me a variety
of photos of me with the glasses and without the glasses because I
trained the model on a sample on a batch of photo that was both with and
without the glasses. This is pretty good. I
would say pretty classy. I like the velvety
kind of jacket blazer. So this is good. Here
are some good ones. For example, this one
looks pretty good to me. Let's see. Yeah.
This little smile there, not quite right. But this one probably
is Let me see. Yeah, this one is pretty cool. This one is very classy as well, kind of like business
entrepreneur. This is a really cool photo. I could use that, definitely. Pretty satisfied about
the result of this model. But you can judge. Like, let me know what you think
of this badge. I have 40 different
ones right here. So casual, some more classical, some business, elegant, I would say casual business with
these ones right here, more like formal business. I like it. I think the model did a great job
understanding my face, my expression, my eyes. I think it's a success. And it is also, in my opinion, better than the previous
model that I trained. Show you maybe the
previous model so you can understand
what I mean. So I'm going to show
you Runny black shirt, so the previous model and go
straight to my headshots. So you see what I was saying
about the mouth being weird. At sometimes, I
trained the model on some photos where I
had some silly faces. And you see, we can see in
that type of photo right here. So, these are the other photos. You see, it's not as good. Like, this one, this guy
right here is not me. I mean, not at all
when you look at that so the photos you choose are really important for training the model and
the model accuracy. Make sure you give it a variety of different photos,
different settings, different backgrounds, well lit, no, like, weird
facial expressions. Something that is
important to understand is what are we actually getting in terms of
image resolution, image dimensions when we
generate headshots app. Well, I've got some information
on this slide right here when we generate
headshots with headshots. We receive high resolution 2048 by 2048 photos. All right. So it's the one by one
squared aspect ratio, but it is in very high quality. So what I suggest you do, okay, once you have taken a good
your 40 free headshots, is that you create a Canva document of
these exact dimensions, 2048 by 2048, and you simply
open the headshots up and you start organizing
your favorite headshot. It's not going to
be the whole 40. Some of them are
going to be failed. So it really depends
on your threshold. Tolerance for how close or not close these photos are
based on who the subject is. But for me, I had a
pretty low tolerance for the errors because I wanted these photos to
really look like me. At least for someone that
knows me from my content, think, Okay, that is running. There's no doubt about
that. All right. So that's what I did here. In this other document, I didn't choose the one by
one aspect ratio, but I did select some of
the best photos here, and not all of them are great. Not all of them are ready
for me to be published. But I selected some
of them because I like the pause or I
like the background. So I decided to keep them
because they will be useful to me to keep
prompting in the future. And we'll see about this
a bit later in the video. All right, so that is important to understand what
is it that you are getting from the headshots app in terms of photo
quality and dimensions. So, guys, we have
trained the model, and we have discovered
our headshots. Now, the next step
would be to generate some new headshots based on the model that
we have trained.
4. Step 3: Creating New AI Headshots: Step number three
is definitely where we start having some
fun with our images. Okay? So now we have
trained our model. We have selected
our best headshots and organized them
in a document. It is now time to
start prompting and generating some new headshots based on whatever we'd like. That is the beauty of
this app and this model. So we are going to start
actually not wild. We are going to start, like, chill and then
progressively go to some more eccentric kind of
photos of me with all right, let me show you the easy way
to generate some new photos. Okay, so here in
the headshots hap, I'm going to come back to the Run model button right
here, the Run Model tab. If you're not there, if
you are on this screen, you will first need to
select your model first. So let's go for the latest
one I train Ronnie. Click on next, and
you will get here. Okay? So from here, you can write a prompt to create a new photo
of Ronnie, right? Because it knows you
now. The model knows. So here is where things
could get tricky, really, because it will
come down to prompt engineering and how
good your prompts are. Okay? So the reason
why I told you, let's start easy, well, is because you have
this button right here, the inspire me button, which is going to do the
heavy lifting for you. So let's click on
it the first time. See as you click on it, it will generate a first prompt. So this will give us a
good indication on how to prompt headshots in order to get good results. So
let's read the prompt. Generate a professional
headshot of male Ronnie at age of 40 with short salt
and pepper hair, brown eyes, and wearing glasses in a
business casual outfit, such as a buttoned down
shirt with rolled up sleeves and tailored trousers with a warm and approachable
expression. The background should be a
modern co working space. Okay, so here repeated
brown eye eyes. I'm not sure why. I
could generate this. Actually, I will generate
this to show you the input and just
know that from here, now every time I will
generate a new photo, it will cost me one credit. When I generate a photo, I actually generate four different versions
of that photo. I told you the magic
media in Canva. Okay, let's generate this
photo and see what we get. I'm not going to
speed up the video. I want to show you
the real speed at which the model works. It says it may take
up to a minute. Usually, it's faster than this. So I think it's a good time for me to have a sip of water. I hope you are enjoying
this content so far. I hope you are seeing
the value in headshots. Me, I was blown away by the quality of the
things I could generate. And also the realism
of these photos. So yeah, I sincerely hope you
are enjoying the content. Alright, my four photos are
ready. Let's have a look. Going to click on
the first one right here and scale that up. Alright, so this one, I can
see not really like me. This face is not me. Too bad because I
like the setting. This one is more like me.
The second one right here. Okay, let me make this one smaller and this
one much bigger. Yes, this is much more
like shirt is not too bad. The background is pretty
interesting as well. So this is something
I could use. Definitely could use this one. The first one, no,
it's a big no. Let's see the third
one right here. This is a pause that
looks familiar to me. Okay, this is pretty
realistic as well. Let's move this one
out of the way. Yeah, this is pretty realistic.
I kind of like this. It looks like a Levi's shirt
without being a evi shirt. I'm not sure which one I
prefer between the two. This one I have a nicer
smile here, I don't smile. And then let's see the
fourth one right here. See if it's any good. There you go it right here. It's not that bad, though. Not as good as the other two. Okay, so I have two
good photos here. Let me scale them
down a little bit, so we can see them side
by side. There you go. So now we have these two
photos that look pretty good. Between the two, I would say
this one is more realistic. So I will get rid of this one and keep this one right here. So, you see, this
was the easy way. I just used the
inspire me button. Could circle through
these prompts and start reading them
as much as I want. Every time I will
push the try another, try another, try another. There will be more prompts, more or less long, and all of these
prompts will work. And they give us a good
template to work with. So from here, we
could go deeper. Okay? So we could start
tweaking these prompts so we can have more
control over the output. For example, let's take
this new prom right here. Generate a professional
headshot of male runnie at age 40 with short salt and
pepper hair hair. Why hair hair every time. Brown eyes eyes. It's weird. And wearing glasses,
framed from the waist up, allowing more of the
body to be visible while still keeping the face
as the primary focus. So from the waist up,
the background should be a high rise office with
a view of city skyline, significantly blurred for
focus on the individual. Capture a professional
expression that conveys confidence
and approachability. Say, frame from the waist up. Ronnie is standing, allowing more of the
body to be visible, still keeping the face
as the primary focus. Okay, let's get rid of that. Ronnie is standing looking
straight at the camera. The background should
be a high rise office. Okay, let's get rid of
the high rise office. Let's just use a
plain background. The background is
an orange gradient. The background is a warm
orange gradient photographer. The background is a warm orange photography
paper background. Ronnie is standing looking straight at the camera in front of a warm orange
photography paper backdrop. That sounds good. Okay, so I'm going
to delete this. The mood is a relaxed expression that conveys confidence
and approachability. Okay, so you see
what I've done here. I've used the base prompt from
the inspire me prompt and I tweaked it to replace
some of the ingredients, some of the elements of that prompt to something
that I would like to see. Now, let's generate this prompt
and see if it's any good. Alright, so this is ready. I have my four different
headshots right here. Some of them are
pretty good, like this first one right here. So this one is pretty good. Let me zoom into this one. Yes, I like the orange backdrop. I can definitely use that. What else do I have?
This one. Not sure. This one I can see, it's not me. This one
looks good, too. So let's have a look, my
traditional black T shirt. I have so many photos
in that black T shirt. So here I have, again,
two good photos. Out of four, this
is really good. Like, 50% success
with these prompts is a really nice
outcome, I believe. So I think you see the
power of this that you can really start playing
around with the real, like, the real character
of you and you can control every single
detail of the shot. Now, I've showed
you some I would say reasonable photos of you. I'm not going to go
into the very crazy, but I want to show you
that you can push this much further and
more specifically, you can use a specific
hack to grab these prompt. Do you see how I struggled writing this prompt right here? It's not easy to write a
good prompt for headshots, because you have to take many different things
into consideration. And I've asked my friend
at headshot, like, what would be a good prompt
for using in headshot. And he told me,
Okay, so the prompt basically needs to have a
few of these ingredients. It needs to have your
subject specification. So detailed physical
attributes, emotional states, and interaction of the
subject with the environment. That would be one, the
subject specification. The second part of
your prompt should be the environmental
context, the environment. So the lighting conditions,
weather effects, spatial relationships, all of these things that kind
of describe the scene. Then three, you should have
some technical parameters. So that could be the camera
settings, the lens type, the post processing effect that have been applied to your photo. And then finally, some
atmospheric qualities. So the mood description, the temporal indicator,
sensory details. So what do we feel when
we look at this photo? So that is kind
of in a nutshell, what your prompt
should be made of, like these four key points. So again, it's not easy to
create a prompt like this. But hey, as always, I've got and guys, I'm going to leave a link to this Canva document with the prompt in the
project description. Okay? So if you want to get
easy access to this prompt, make sure you check the link
in the project description. This will lead you
directly to this page right here where you
can copy my prompt.
5. Bonus: Use Photos as References!: Alright, guys, in
this bonus section, I'm going to show you my favorite way of
using this app, Okay? Because the inspire
me button is nice. Writing your own
prompt is complicated. So there is an actual hack to
do it in a much easier way. And that technique consists of using an existing
photo as a reference, and to use Chat GPT to describe that photo
to us and turn it into a prompt structure based on the prompt structure
I mentioned before. I'm referring to this prompt
structure right here on this slide that I read to
you just a few minutes ago. All right. So let me show
you how we would do this. The first step would be to find an image that we kind
of want to replicate, a pause, a portrait, an attitude that
you want to mimic. But with you as the Alright? So me, obviously, I research famous
entrepreneur portraits, and very quickly, I stumble
upon Steve Jobs, right? And this famous the
very famous portrait of Steve Jobs, kind of, like, pulling his chin hair with his black turtanneck
iconic sweatshirt. Okay, so let's see if
I can reproduce this. I also very much like this one. This photo of Steve Jobs, pointing, like, a
very determined look. So let's see what
we can do here. And then I will show you some of the other project I've
worked on right here, and you will see
it's pretty playful, pretty fun, pretty
colorful, as well. Alright. So how would that work? Well, first thing first,
you need the photo. Okay? So, this one, I grabbed it from Google, but you can grab a photo in the elements tab right
here. You can search. I'm going to show you some
of my recent searches. So this lady right here, I just typed Astronaut portrait. I think that's what I tapped for finding this
photo right here, and I'll show you
in a minute how I reproduced it with my model. So you can search the
Canva library for some photos by simply using
the elements tab to see. So I had Mexican wrestler
that I tried as well. Need to go to your photo
section right here, see all, and you will
have all photos. Okay? So I try one of these, so this set right here. I also show you the
result in a minute. Anyways, once you have this, you can go to HAGPT and use
this exact prompt right here. Okay? So you can screenshot
this slide because this will be your base
basically, it goes like this. Analyze the essential components of the photo attached
and help me generate a new technical prompt
for a character called Runny that would match the
attached photo exactly. The prompt for Runny
should be no longer than 850 characters and needs to include,
then I put the three, four sections of the prompt,
subject specification, subject outfit description,
environmental context, and technical parameters, and also the
atmospheric quality. So there are five parts to
this prompt right here. Okay, so copy this prompt, and then head over HGPT. You can start a
new conversation. GPT 40 will do. You just need to be
able to attach a file. Now, I believe even free
users can attach files now. You don't even need to use HAGT. You can use any
large language model that allows you to
upload a file, a photo. So it would go like this. So first, paste your prompt
and then attach the photo. I think it's this one
right here. There you go. I so now you run the prompt, and you see Chat GPT is going
to come up with the prompt. Okay, runny, middle
aged man with a confident yet
contemplative expression stands in a striking
close up portrait against a seamless
white background. His right hand gently
rest under his chin, fingers curled slightly, conveying deep thought.
Blah, blah, blah. Okay, so you see the image is captured with a 50
millimeter lens f 2.8 for sharpness with
monochrome post processing, adding a crisp, yet
organic film quality. It is pretty good, though
it has been bolded here. So I will ask to remove any
formatting from this prompt. Okay, not changing the prompt, I just need it to
remove the formatting. Okay, so when I paste
that into headshots, I don't have the asterisks that are typical of the
bolded text in a prompt. Just going to copy it and come back to my Canva
document right here, go to my headshots. I'm
going to clear this. I'm going to come to this
page and paste my prompt. Now, I see that there is a few things that are mentioned
here about the glasses. I don't want these glasses, but I want my glasses. So let's put close up
portrait of Ronnie. Okay. Basically, I just get rid of what is
not necessary here. Close up portrait of Ronnie with a confident yet
contemplative expression, stands intra his intense gaze, framed by his glasses. Okay, I think that's
good enough now. Now, I can go ahead and generate this for one
credit. Let's see. Alright, guys, my four
Steve Jobs photos here, looking at them, I can see immediately that
this one is not going to cut it too bad because the position of the hand was pretty good. This one could work. Let's see. This one is not too bad, though, the beautifier went
crazy on this one, I think. But it's not too bad. Let's see this other
one right here, see if it's a little
bit less beautified. I like the second hand here. This one is nice, actually. So you see, I was a to reproduce something pretty
close to the Tepjob photos. But yeah, if you're not
happy with the result, you can keep tweaking your prompt or you
can generate again. Like, you can simply
generate again. You will get four new photos, and AI is a little
bit like this. So it's a probability
game, for sure. You can tweak your
prompt or you want. It's still going to
be an iteration game. So the more you
prompt, obviously, the more chances you get to
having the perfect photo. Alright, I think I'm
going to leave it here, but before I wrap
up this video I want to show you some of the other experiment
I have been running. They are right here
in my other document. So let me get to this page right here and go
full screen so you can see. Remember, I told you I used this reference image
from the Canva Library. Okay, so I search for this
one in the alva Library. The terms portrait astronaut. And then I used the
same technique. I put this into HAT GPT,
used the same prompt. I got a description that got
me this photo right here. And this one was with the
first model I trained, the Ronnie black shirt model, not even the best
model. I like this. I think it captured the essence of this
photo pretty well. So lost the helmet.
Try to get the helmet. I didn't work. This one
from this guy right here, generated this one, which I
think was pretty good, too. Very realistic, kind of scary. I like the texture of
the wrestlers mask. This one was pretty
good. Tried it with some other colors. So this one to go from
this one to this one, the only thing I changed was the description in the prompt
of the colors of the mask. So I just changed the three colors for
three different colors. This one, kind of weird, but from this photo right here. So this one more for maybe an album cover or I don't know. Spotify playlist. I tried to replicate
this effect of the face kind of pushed
against the plastic. Didn't really render that much, but I thought it was interesting and I thought
I would show it to you. This one really like this one is cute based on that image and the prompt description
that I got from Cha GPT could generate
this photo right here. I just changed the
description of the glasses for just
my original glasses. So this one pretty cool, too.
6. Class Project: Congratulations, guys. You made it to the
end of the class. And now it is time
for those of you who want to participate
in the class project. And you would have guessed
it the project consists of creating your very own batch
of photos with headshots. So you will need to go through all the steps that I
explained in the class. You'll have to first
train a new model, discover your
existing headshots, and then play around with the prompt box to generate
your own headshot. I don't want to see
the generic headshots. I want to see some
crazy one. All right? So once you do have a few good ones that
you are happy with, by all means, post them in
the project description. Explain to us, maybe you
can add the prompt that you use if you are into
this. Give me some context. You can also use the
technique I show at the end where you use
a reference photo. So in that case, you show me the reference photo
and then your photo. That would be a cool way
of using the project. I really look forward to
discovering your photos. I will jump in there and
give you my feedback. Definitely, I always do that. So yeah, get creative
and surprise me.
