Do you want to generate images with your face on them. ? Then this article is for you.
So, i generated some with this prompt ‘Matt as a gorilla human hybrid,’ and apparently, Matt as a gorilla human hybrid is just Dave Batista. Here are some with Matt as a bodybuilder flexing muscles upper body.
Real Images vs AI Generated
Now, you might be fooled into thinking these are AI, but these are actually real images. Just ignore the fact that there’s a prompt, and it looks like they were AI generated. These ones are totally legit.
Simplifying the Process with Artflow
Now, in the past, to get images like this, you had to use a crazy complicated Google collab, where you adjust a whole bunch of settings and tweak a bunch of stuff and upload a whole bunch of training images and decide how many Epoch you want. It’s a pretty intense process.
However, I made these images with a tool called “Artflow” which takes all of the confusion and complicated steps out of using something like Google collab.
Artflow: Streamlining Character Creation
Now, Artflow, which you can find over at app.artflow.ai is actually designed to create consistent characters. If you’ve played around a lot with Mid Journey or Dolly or tools like that, it is ridiculously hard to get any sort of consistency in the characters you create.
Now, stable diffusion is capable of doing that, and from what I can tell, Artflow is using stable diffusion underneath the hood. It’s also most likely using dream Booth, the same exact technique that I used to train my other likenesses into the AI under the hood as well.
But once you create a consistent character, you can then generate images with that consistent character and even create videos with that consistent character. Training your face with Artflow itself isn’t necessarily free.
Usage and Plans
They do have a free plan that gets you 100 credits a month, a handful of minutes in video, and it does actually have watermarks if you’re using the free plan. Then they have other plans that give you more credits.
However, from what I can tell, one credit gives you four image generations. So you give it a prompt, you get four images for that prompt, and that will use one credit. So 100 credits per month is roughly 400 image generations per month.
Creating a Character with Artflow
Also, to train your face into the AI, it’s actually free for your first training. So if you just want to train yourself, it’s free to do that. So I’m going to sign up for a free account real quick, and once I’m logged in, I’m going to click on character builder up in the left menu here.
Here we’ll go ahead and press the button to create a new character. It asks me to define the character. To start, I’ll just put “white middle-age male with dark brown hair and a beard.” Then we have the “Add consistent face” button, and I can click “Choose face.” We have a whole bunch of existing faces and characters to choose from. Let’s go back to actors and click “Train actor.” This is where we could train our own face into the platform.
Additional Features
Here, I’ll click “Create your AI actor.” I’ll enter “male actor,” “adult actor,” and input your name “.”Then upload 20 photos of your face and click “Start training.” It might take 10 to 30 minutes to process. We’ll check back shortly.
When using the Google Collab Dream Booth method, you had to leave the browser open, scroll around, and make sure the page didn’t time out. With Artflow, it seems like you can step away and check back when it’s ready. The training takes about 15 minutes.
If you try to create another AI character, it will asks you to upgrade, but the first training was free. Clicking “Create image” next to your face, generated a bunch of demo images. These include images as a wizard, the Joker, and an invincible superhero.
Export and Animation with Artflow
In the Animation Studio, you can enter some dialogue. Like “Subscribe to “” on YouTube.”
If you click export and animate, it will use two credits to make the short clip. After a few seconds, you’ll get a video that looks like it: Now, it’s not perfect, but you do have the option to upload your audio file, so you could record an audio file of yourself speaking or create an audio file from 11 Labs that’s trained on your voice. Pull it into the upload section, and you can make it sound like you.
Artflow’s Impressive Capabilities
What impresses me most about this Artflow AI tool is the ability to quickly train your face and generate any images you can imagine, all trained on your own face. Let’s try one more thing.
Experimenting with Prompts
This time leave director mode off. put prompt “Matt Wolf as Superman flying above a city skyline.” It made an attempt, and as we know with AI art, sometimes it takes a few prompts to get exactly what you’re looking for.
Here’s the result : it thinks my face looks like the Hulk, as a newscaster (part of my thumbnail strategy), and as a Viking Warrior.
I can’t get it as dialed in and perfect as with Dream Booth and stable diffusion using something like automatic 11-11 or Comfy UI. Still, I can generate images that look like my face on whatever subject I’m trying to put them on, and it’s probably ten times faster than using Dream Booth.
Final Thoughts on Artflow AI
In conclusion, if you just need a couple of images here and there with your face trained into the AI, Artflow AI is perfect for that.
You can pretty much do everything you need for free unless you need to train more than one model or you’re planning on generating more than 400 images a month. In that case, it’s probably worth paying the extra few bucks and being able to create a lot more generations.
