Відео

I wanted to honor it, and tried to make it as good as I could from the line art alone. Thanks!

@reallybigname great job mate 👍

😂😂😂😂😂🎉❤

If you like the hair stuff, I have a whole other video with more! haha ua-cam.com/video/zfyms8XwNZw/v-deo.html

This is just what it looks like when I experiment with software... CogVideoX specifically.

Oh, and I'm in the process of making a lot of music for albums. The music is one of the rejects from that process. I mean, it's good, but I have better!

@@reallybignameI appreciate you copping such harsh feedback haha. where can I listen?

@@fnytnqsladcgqlefzcqxlzlcgj9220 you can't yet. I will definitely promote albums when I have them! 've been preparing to get a Suno membership for a long time, practicing my prompting on free account. But, I finally just finally paid for Suno like a week ago! So far, I have about 150 songs in a bucket that I marked as ready for an album. (I generated a lot more than that) I'm about to sort them all out into different albums buckets, each different styles. Many songs have multiple versions in different styles as well. So, I might also make some albums that are all the same song, done a bunch of ways, in different eras, and I'll have fake backstories and fake band names for them. Anyway, the next step will be to figure out how I should get them out there for sale. Any suggestions? Just never done it. I was just going to look up the available services that can get stuff on the main platforms and ship physical albums. Then there's promotion, etc... They have this brand new feature for doing "covers". If uploading, you can upload a minute of the song you want to cover. All your melody and general structure has to be there in that minute. But, I've gotten it to do some great things to my old songs using that minute. See, I was already a musician and singer/songwriter, so I have tons of stuff to draw from. Also, I can also just sing whatever I want into a mic and then turn it into something else. It is such a great feature so far! I've made some of my songs sound like Freddy Mercury and shit, but they are singing my melody, my trills, etc... it's so trippy. It is crazy stuff that you can't do with just the base prompting too. And, you can also cover a song you made with Suno, and if you do that pays attention to about 3 mins of the original song, rather than just the uploaded minute. It will fall off the melody and start being monotonous (you can tell pretty quick). But,, then you just use the Extend feature and extend from where it got stupid. It generally does a good job from extending an established song. Doing originals is great now, but this Cover feature has also made Suno-generated songs so much more dynamic, since you can just keep doing covers of covers, changing lyrics, changing styles, etc...

@@fnytnqsladcgqlefzcqxlzlcgj9220 Oh, but you can see tons of my free Suno stuff here: suno.com/@reallybigname

I thought you needed 12GB, but you should check the CogVideoX github. I dunno. I have 24GB.

Thanks! We finally got a decent video model (CogVideoX) that you can set up locally. But, it can only do 49 frames at a time, so no continuous video possible yet. But, I can do stuff like this! I set up a prompt with random elements that would change, and auto-generated like twice this many, then tossed out the worse ones to make this vid.

With the latest NF4 model, I can do 1024x1024 at 20 steps in 13 seconds! Flux.1 is fun!

Thanks. All I did was barely blend each clip over like a quarter second. LumaLabs just makes great clips! But, to avoid a stall, you do have to blend a little tiny bit.

@mukejagger8503 thanks! Yeah, I got depth working with affine and perspective motion, but I still have a bunch to do to get everything playing nicely together. But basically, instead of choosing either optical flow or one of those modes, I think I'll be able to let users use both! So, you could do flow plus perspective motion, or flow plus affine, and the depth of the image is used during the matrix transformations. And, in trying to make another mode of my own that can do something kind of in between, like perspective, but canceling out the skew. Now I just need to repair the rest of the stuff that I broke and redo the ui for hybrid motion! Ugh...

@mukejagger8503 oh, and also working on an accurate z translation estimate from video. Those matrices can get the angle of the camera, but not the front back movement. If I can get that working and aligned, you'll be able to turn compositing off and just do your own animation, but using a motion path from video! That's the ultimate goal.

Glad you liked it! Sometimes I just see something and I have to execute the plan!

@@reallybigname I took the video of the first Crew Dragon docking and put it to Star Trek the Motion Picture music when Scotty was piloting Captain Kirk to the Enterprise :D I get the feeling!

@@LTeppler Sweet!

Here were the settings I used (Deforum Discord on the webui-reforum channel: discord.com/channels/1011314943380959313/1232929179805421629/1247308478201069639

@@reallybigname This link seems broken? I can't seem to find the webui-reforum channel on Discord. Can you tell me the model you used? I can prolly figure out the rest.

@matthewbennett link worked fine for me. Maybe you have to add it to your channels list.

@reallybigname i think I need an invite to the discord channel? I tried on my phone and computer, both logged into Fiscord, and the link takes me to a dead page.

@@matthewbennett Try this discord.gg/MJVcPfrN

Settings for this one and others are posted on the webui-reforum channel on the deforum discord. And, I'll be posting more there. I do want to do an overview video soon too though.

@@reallybigname that’s awesome! I’ll definitely check it out! Amazing work 💪🏼 maybe we should call you reallybigbrain instead

Thanks! I wanted to show off some of what my Reforum branch can do. I may do a tutorial for people soon. You can read about it here: github.com/reallybigname/sd-webui-deforum/tree/reforum