THIS Changes AI Video / Metea EMU uses your Data? AI NEWS

as always there are more AI news and here is what you need to know of course I'm still in Bangkok absolutely enjoying my time here with the cuteness on every corner I'm a huge foodie and this is my double chocolate shake for only $2 this time on Central World the Christmas decoration is Pokemon with hundreds of Pikachus on every corner of course Bangkok also has amazing skylines there's tons of skylines everywhere in the city some more food I en Choy this complete dinner was only seven bucks and then one more Buck for Endless green tea and these one piece tumblers only cost $3 a piece I think I need to get one let me know in the comments which one I should get and I found this questionable piece of marketing but I take it with Good Humor the first news for today is the new Facebook image generator called emu and Blain Brown has some pretty cool images here we have a ice dragon in kind of a Christmas scenario I really like how the colors look and the composition everything looks very good for a newly released model so this is very impressive of what Facebook has presented here and I find everything about these images very stunning however there's also another bit of information meta used 1.1 billion Instagram and Facebook photos that are publicly available which means user data so whatever you have uploaded to Facebook and Instagram could have been used to train these IM Imes and that's kind of interesting especially from a privacy issue because these are not artist works these are private images of course they're shared publicly but still they are from individuals not from professionals but let's have a look at the video from the Facebook event before today you know we haven't talked that much about our image generation models now we call it emu today we're starting to roll out a bunch of products with with the Senate and you know it's they're high quality photo realistic but you know one of the coolest things is that emu generates them fast right so it's not a minute these are it's about 5 Seconds to generate one of these so you can just message it and tag anything with imagine and it will generate highquality photorealistic images right there in the chat in seconds for free the next bit of news is very important and very close to my heart event by civid AI it's called get lit and give back now it's a little bit hidden here in the announcement but actually this is about juvenile diabetes research so everything that you donate in the course of this event will be matched by CI as the amount and then be given to this foundation for the research a very nice Christmas guesture and a very important event I hope you take part in that and the next bit of information is also about civi the dream shaper 8 LCM model is out you can see here here super nice quality you can render an image in 1 second flat with the new LCM technology that I showed you before in this video and as you can see here the results are absolutely stunning and high quality you can see here some really amazing images they also have a very nice variety of different characters and materials here you have a little bit of a mix between digital art and a water painting effect and even anime can be rendered because dream shaper is a model that can do a huge variety of different art styles so if you want to experiment with different or even with different style presets dream shaper is actually a very good model to experiment with that next let's talk about loose control lifting control net for generalized depth conditioning now what this is doing is very impressive a very interesting concept as you can see here you can use very basic shapes like just 3D boxes and use that as a live input to render with stable diffusion and get a live feedback of these changes that are happening as you can see here this has a huge variety of different applications and I would say this is also a fantastic outlook on what can happen when we have ai that is actually generating on live input so we can imagine this as an input for movies or games that are rendering the scene based on a live input that the scene is providing as basic shapes this of course also makes it much easier for you to actually create film scenes where you have full control over the position and the movement of different objects and characters inside of that scene they have a lot more examples on the website on how this can be used by these different previews here a lot of that has to do with different room compositions but of course as you have seen also you can manipulate objects and move them around in the scene the cool thing is that you can also try this today live you can see here up there it's a prompt input see input and down here you can manipulate this kind of 3D shape you can make it wider and you can also adjust the camera to get closer to that you can rotate the camera everything here is a little bit clunky and also the rendering down here is very slow because a lot of people are using it but it gives you a chance to experiment with this at least a little bit and there is even more fantastic news for AI video generation this project is called motion control and as we can see here in the examples this enables us to control the camera movement angle and rotation inside of the AI video bringing us much closer to becoming AI video directors and another fantastic thing you can see in the lower examples is that this also allows for the control of objects inside of the video the interesting thing here is that the objects even follow the logic of how they are supposed to work so you can see that the the wind charm on the right side is actually dangling in the air while the sunflower on the left is slowly and softly swaying in the air when we scroll down on the page we can find a lot more other examples and the interesting part here at the lower examples is that you can actually control multiple characters in the same scene so this can be over time interesting for controlling different characters or actors or even crowds inside of your scene of course this can also be used for more classic things like panning zooming or even rotation of the camera in different directions but we're not done yet with the news about AI video generation so here is a video from Nix Pier about Pika 1.0 and you should try that out on their Discord server really amazing and all of these fantastic scenes have been created with paa 1.0 so it's really cool how realistic they look what how filmic they are and also the consistency of the scenes is very impressive with the details in the face with the Cinematic expression of what we see in the scene really really stunning now let's also check out some fun and Innovative Community projects here we have the SD Christmas Card Factory uploaded by neural L you can download that comu ey workflow for free and this is hosted of course on the open Art website which is quickly becoming the database for comu ey workflows you should check it out not just because there is still the contest going with a price pool of over $113,000 but also because you can find here many many workflows for all things from image generation to IP adapter to video generation to animate diff there's a lot of different things to explore here and all of these workflows can be downloaded for free next we have a really cool project here from a Reddit user he combines SD Turbo with stable fast to create 149 images per second with stable diffusion now that is crazy next we have here a fantastic project by another Reddit user who is doing realtime imageo image translation you can see the realtime video input and then the output that is using touch designer text to image adapter Kenny sdxl and the Turbo Laura very impressive technology very Innovative use of stable diffusion and of course you should absolutely check out this very Innovative video here it's using comfyi and Anime diff so really cool result here very smooth motion I'm absolutely impressed by this video let me know in the comments where you see the future of video AI thanks for watching and leave a like if you enjoyed this video see you soon bye oh you're still here so uh This is the End screen there's other stuff you can watch like this or that's really cool and yeah I hope I see you soon uh leave a like if you haven't yet and well um yeah

Channel: Olivio Sarikas

Views: 27,169

Keywords: oliviosarikas, olivio sarikas, olivio tutorials, meta facebook, meta emu, meta ai, facebook emu, facebook ai, ai news, best ai tools, new ai tools, ai future, ai video, OpenArt, ai money, ai make money, stable diffusion, comfy ui, automatic 1111, midjounrey, stable diffusion ai, ai video generator, a1111 stable diffusion, stable diffusion xl, lcm, sdxl turbo, svd, svd diffusion, svd video, animatediff, comfyui, pika labs, animatediff comfyui, midjourney, automatic1111, sdxl

Id: ozQ0LloTfP8

Length: 9min 10sec (550 seconds)

Published: Thu Dec 07 2023