[AI Image Generation] New Model 'Tsubaki' (Beta) Released! What are my impressions!? : PixAI🐳

Thank you for finding this article🥰

I'm Natsume, and I want to walk alongside AI while using the AI that evolves every day😊

This time, a new Beta version model has been released by PixAI, so I tried creating with it🥰

Its name is

'Tsubaki'

It's a model with a 'Beta' banner on it🥰

Since it is not an official release but a Beta version,

I couldn't use LoRA🤤

Hmm?
Are you telling me, Natsume, a near-beginner at PixAI, to draw without the power of LoRA?
(I'm super anxious ( ^ω^ ) ...)

I have never been able to make the most of PixAI without LoRA (´;ω;｀)
I guess that means my prompt construction skills are poor (´;ω;｀)

( ^ω^ ) ...
(Super anxious)

I won't know until I try!

This time, I put together a prompt with a 'girl drawing the universe' taste🥰

Well then, let's give it a try!

For comparison, first I will create with an existing model without and with LoRA.

First, I tried creating with an existing model without LoRA🥰

[Existing model, no LoRA]

Can you see that I am

not making the most of

this image generation AI?
To put it into words,

the illustration is flat

and lacks depth.
It's 'not capturing the universe,' is it?🥰

This is due to Natsume's lack of technical skill; I must strongly emphasize here that the model itself is top-tier!!!

Yes, let me show you what happens when you add 'LoRA' to this🥰

[Existing model, with LoRA] *Same prompt

Could you see how the depiction changed dramatically when I 'customized' the 'model' with LoRA magic??

When you think about the depiction, LoRA magic is amazing, isn't it?
If you can find the perfect combination by utilizing the original model and adding LoRA, then finally...

you become a master

of it, I suppose ( ^ω^ )...✨ (looking into the distance)

So, here is the new main model for this time,

'Tsubaki' (Beta model, no LoRA)

and of course, if I have it draw with the same prompt, what will happen...

I tried it out🥰

( ^ω^ )...??? Wait?

What on earth could this mean???

Is this what it's like without LoRA???(●´ω｀●)💦

The 'Tsubaki' NEW model is truly formidable...💦

Thoughts: LoRA... is it... necessary??

Judging from the prompt, I really felt the AI's privilege of

'creativity'

in it.
For me, what I look for in image generation AI is

depictions that humans wouldn't think of or come up with

on their own.

To put it negatively, that is not 'following the prompt.'
But I want the AI to 'freely imagine' the prompt and

create something.

With that thought in mind, I intentionally avoid putting together overly restrictive prompts🥰

I want to create 'creativity' between humans (realistic) and AI (imaginative).

I felt something close to that in this model🥰

If it gets an official release and we can add LoRA, what kind of world will be waiting for us???

...This is, gulp ( ^ω^ )..., I'm looking forward to it✨

That's all from PixAI🐳

That's it for this time🥰

Thank you for reading this far🥰

Thank you always for the likes, comments, and adding to your magazines (´;ω;｀)✨

Please continue to support me🥰

Well then, have a great AI image generation life~💐

🔽PixAI Related Works🔽