Join me on Facebook: facebook.com/groups/theairevolution Join me on Discord: discord.gg/XKAk7GUzAW #### Links from the Video #### sites.google.com/view/stylegan-t/ www.reddit.com/r/StableDiffusion/comments/11m280v/fantasyai_claims_exclusive_rights_to_models_that/ www.reddit.com/r/StableDiffusion/comments/11mrzzq/fantasy_ai_statement_here_to_answer_your_questions/ github.com/Sanster/tldream xanthius.itch.io/multi-frame-rendering-for-stablediffusion huggingface.co/thibaud/controlnet-sd21 github.com/microsoft/visual-chatgpt

Thanks so much for making these great videos and always keeping us to date. It's often difficult to keep track of all the latest development (especially the more useful ones). Thankfully you're always there to offer insight and clarification!

I feel like we need an AI to keep releasing daily videos about AI image generating. Love seeing these rapid advancements in this technology. Thank you Olivio for these vids!

The love and addiction for AI tools is the main thing that keeps me active these days. Thank you for keeping us updated Olivio.❤🤝

Wow! ...especially the SD"animation script" seems to be a promising further development. Big Fanx for sharing!

Thank you for supplying the newest information about image generators!! 👍🏻

Imagine building a small area manually, or in whole or in part with AI, and then being able to have it imagine the rest of the world based the assets and style and story you’ve set up already. It just generates an endless amount of content just over the horizon. Amazing

Thanks for your SD guide! :Liked and subscribed)

3:28 That animation, is crazy! Hold onto your papers people! Hold onto them tight!!!

I really enjoyed this episode, so much coming out in the Ai world. Amazing.

this is probably a stupid question but do any of the web UI programs support StyleGan models?

Thanks for your excellent review. Do you know wich text to image generates larger images natively, without using upscale?

Thanks so much for the vid! So exciting

I think comparing speed of a diffusion based system to a GAN is a bit disingenuous as they are really completely different things, and generate very different results. GANs were what we were using before SD smashed into the world and I don't think anyone wants to go back to GANs :)

Thanks Olivio! I enjoy all your videos, but I have a problem, and I might need some help. I want to hop on the train of AI image/video generation so badly, but there are new tools coming out left and right every day, and it's getting quite difficult (for me) to keep up. I just don't know where to start! I feel like at this point, it would be a good idea to make a quick Recap Video and show the best way to get started from scratch. I know you've already made videos in the past about that, and so have other creators, but the rate of evolution and progress of all these tools is insane! Everything is moving so fast! For a complete beginner like me, I would love to see an Updated guide/video showing me what to download first, where to download it, how to set it up, which tools/add-ons are the most popular/useful, etc. Some people might say "Just pick one and experiment. It's easy.", but something I have struggled with all my life is *Taking Action*. Whenever I feel overwhelmed, it becomes harder and harder for me to actually take that step forward and *Take Action*. I've been contemplating AI image generation for months now, and look at me: I still haven't hoped on the train. I am, frankly, getting overwhelmed by all of this. I know this is a personal struggle, and I hope people don't come at me to ridicule me or make me feel worse about myself (This is the internet after all, and it's often filled with toxicity). I know there are other people out there who might have similar struggles as me, so I'm taking a chance here to reach out. I just need a little push, and I know that once I actually get started, I'll be unstoppable! Thanks!

Great video. Nowadays, anyone can utilize AI to quickly, simply, and economically produce great images. I recently utilised Blue Willow, and it really awesome

Thanks for content as always, just a suggestion, maybe add "AI NEWS" or something to the title? I was excited to think this video was solely about this new method but it's more of a general news video? Which i think is great anyway as there is so much stuff coming out that it is easy to miss something but if it's going to regular maybe note the video as News? Thanks again for all your content. :)

Quick AI news. Great, love it!

Nice news format!

I love your channel!!. How do you generate the toon character at the end of your video?

name suggestion: AI_standing news with Olivio :> or ask chatGPT about suggestions. Didn't know that you doing such new feed - I like it.

Love the news content. Quick question - Where do you get your news? Things happen so fast with image generation it's hard to keep up!

Thank you for making my coffee lazy moments so awesome.

Hi Love thanks for all the wonder videos :) can you please put a link to these sites ? I am trying to go to style gan t .

Do you recommend installing stable diffusion 2.1?

I was laughing hard at the "for the v2.1 fans...is there any?" I think I heard crickets when you asked that question.

Super, extra fast. Sounds like "double, secret probation". But I am looking forward to my web renders taking less than 0.4 seconds. That's faster than I can change my prompts. Of course, 3.7 seconds is faster than I can think about what I want to DO to change my prompts. :-)

Great video :D

Looks like Civit is down actually. Anywhere else those files are hosted?

AI changes so quickly you can't even finish a video before there is new news lol. Thanks for sharing this info though!

Does anyone know an article or blog providing a complete controlNet guide?

so the Standart control net models dont work with sd2.1 models? why is that?

Woah, you should try that IMG to Video. Would love to see from your perspective

Control net for 2.1 is huge

How to install in windows this Visual ChatGPT?

StyleGAN-T already got blown out of the water by Adobe's GigaGAN. Now that we know GANs can be made at least good enough, I think they'll retain a place in creative work because of the clean latent space, even if diffusion or transformer models stay ahead in quality.

Thanks.

So now we can't use those models on our own now? It has to go through their site!?

can you explain how to install visual chatgpt please?

Hmm... *die Liste der Modelle anseh und grübel* Gut, ich bin noch lange nicht an dem Punkt wo ich Favoriten habe weil es so viele Modelle gibt und ich nicht genug Zeit habe um wirklich zu experimentieren. Aber... ich habe von allen genannten Modellen bzw. Modell-Creatoren mit mindestens 1 Modell getestet und die sind eindeutig alle in den Top Kandidaten für eventuelle Favoriten. Aktuell mache ich mir keine wirklichen Sorgen... 14 Tage Exklusiv-Download... ich hinke eh hinterher und habe kein großes Interesse auf Webseiten zu generieren. Ich hoffe das bleibt so in der Richtung...

There was a easterly breeze 5 days ago

one step closer to VR live '')

can you makre a video on how to instal visual chat gpt ?

mindblowing. A bit angry about the fantasy ai thing.

There's also new finetuning method that takes only 0.5s and seems to be pretty good quality. It's called ELITE

Anyone noticed that stable diffusion model names all sound like weed strains

I don't understand why so many videos talk about Google's amazing AI breakthroughs, if none of these things are usable. They are not open source, they are not available to anyone, you can't even try them. Where can I enter the prompt and get the images? Where can I enter the prompt and get the music? So many papers, so much blah blah. Of substance? ZERO! Google is making itself more ridiculous every day. It lost the AI battle even before deploying its first warrior.

Are any of these generators available for use on our own machines? Stable Diffusion is anything BUT stable. It keeps breaking, and I have to keep reloading it. I'd love to have another option.

At this rate, we will have full animation capability in AI within at most 2 years.

I'm not trying out a AI image generator that's has a trial period

10gb for one model 21, maybe link on smallest?)

Yea they're all merges, and they haven't taken Neverendingdream - and AI shouldn't be LISC like that, donations are one thing - but lisc? XD That's like me asking people to pay for LoRa on art that i haven't made, and to be honest? That would include things iv'e already shared in your fb group - Paywalling anything that isn't my own art is borderline illegal in the ethics thing XD

I feel like I should get a 2 TB drive just for the models...

You should add a mention that this is an AD.

Quality > Speed

Realistic Vision is merged using 10 different models. Did Fantasy AI pay ALL those creators? And they say money they get will be shared with original creators, I doubt it goes to all the creators who made models that were merged into Realistic Vision and others. Luckily, I don't use any of the models on Fantasy AI. But who knows if they or another company will continue to buy out model licenses. Very concerning.

no code yet

GAN is like 7 years old method...

first one to comment😊

Yes! I'm a fan of 2.1, I don't know why people ignore it so much, the negative prompts work great. Sure there are not many models for 2.1 but it all depends on what your specific use of SD is for.

make a video how to use this speedy ai

This kind of license is like a disease.

Olivio you make great videos and you're the only ai tech person I regularly watch, but you gotta stop using the weird poorly animated cat thing. It's completely out of place.

Kind of misleading to not show the results of StyleGAN-T. It's half a decade behind stable diffusion in terms of the quality of images that stable diffusion can produce. It doesn't matter if it's 30 times faster if the results are 30 times worse.

Yeah. But it’s at 256x256. Why is no one mentioning this. Of course it’s going to be much much faster. And the quality doesn’t look good either.

10 images a day for free lols hard. Im doing 300 a day at 1024x1024 then up scaling it to 4 and 8 k for free. are these people losing their minds?

You say style gan t is faster "for the same image". It's not the same image. Look at the fid column. The score is really high for style gan t. High is bad. The reason ediff-I is slow is because it's really good. Style gan t is fast and bad

why? why always become paid subs....................

So now you're getting mad because someone uses your models as they please and make money off of it? Remember how you used billions of people's artworks as you please and make money with it and felt no one needed to be compensated for that? Just a reminder it, it was AI users who were all in for this wild west attitude where everything is up for grabs and where you can use anyone's work in your programs. And now they are surprised and crying because someone is treating them with that same attitude lol.

Cant watch right now, but you better not promote fantasyai, or the community will boycot you too, we are already talking about getting them banned from reddit

Less of these old "news". This is waste of time with fake title. Make video when that is released and when you are using that. That "news" was released weeks ago.

civitai or civit-ai? 🤔