Let's Take a Look at Stable Diffusion's Next Big Thing, SDXL- Dates, Stunning Results, Try it Now!

Поділитися
Вставка
  • Опубліковано 28 жов 2024

КОМЕНТАРІ • 15

  • @DDCRExposed
    @DDCRExposed Рік тому +2

    Relight is great as it appears the size of the light indicates a Z-Axis in the image. Can't wait to see A1111's adaption of this.

  • @chrisvowell2890
    @chrisvowell2890 Рік тому

    I'm a total newbee to AI graphics and this program has blown me away. From short, plain language text prompts, I've been able to create some truly stunning images. One of the things I've noticed which is disappointing is multi fingered hands and extra buttons on clothes but I guess this will be controllable in the full release. I've looked at other vids to get clues how to use AI (particularly on these problems) but they've all referred to earlier versons which are notoriously difficult to install and then to learn how to use! Hopefully the full SDXL 1.0 release will be easier to install.
    (Incidentally, I've been cropping my copies of full images to remove the watermark or used psp to clone them out, but I will definitely try the 'clean up' routine. I didn't hear it from you!)
    Well done - very useful. Subscribed!!

  • @musicandhappinessbyjo795
    @musicandhappinessbyjo795 Рік тому +1

    Ok thanks for the advice about not cleaning the water mark.

  • @mikesrandomchannel
    @mikesrandomchannel Рік тому +1

    I’m just looking at this and saying: Let me in/out paint with it. Which is why I’m mostly interested in in/outpainting with SD and also PS with gen fill.

  • @bladechild2449
    @bladechild2449 Рік тому

    man that clipdrop was awesome 2 days ago, took maybe 30 seconds for prompts. Now it takes like 5 minutes lol

  • @sasanpourmasoumi5256
    @sasanpourmasoumi5256 Рік тому +1

    Thank you so much for laying out this info. As an artist and animator who has been using rather traditional methods creating art, I still often find it frustrating not to be able to do small tweaks or draw some details first hand while using these apps. Lol. I guess it would be great to combine these prompt based tools or integrate them into a pipeline where users could go back and forth from hand drawing/ human editing to prompts. or maybe there is such thing and I'm not aware of it. great videos btw

    • @ArtificallyIntelligentAi
      @ArtificallyIntelligentAi  Рік тому +1

      Hi there! no problem at all. I'm glad you found the videos useful. I wonder if PhotoPea extension can do what you want. I did 2 tutorials on it. It's essentially a free version of photoshop that you can use right inside stable diffusion and send the photo back and fourth from photopea to control net, img2img etc.. The reason I suggest it is because you can do layering inside the app and grab some textures and colors from the original img to draw with if you have one of those drawing pads. You can also use chrome desktop to remote in via your tablet and use the stylus to draw and process everything on your PC from your tablet. Here is my first photopea video ua-cam.com/video/U_T5qHSwY8E/v-deo.html the second one should be in the video description if its what your looking for. The app does have ads but PhotoPeta has been around since 2013 and the adds come in through an iFrame from their actual website. It works like an embedding. I'll also do a remote desktop video since it seems like a cool idea to be able to use A1111 from any PC, tablet or phone.

    • @sasanpourmasoumi5256
      @sasanpourmasoumi5256 Рік тому

      @@ArtificallyIntelligentAi wow. Can't wait to give it a try. Could send you the result. I think a comparison of result( output, pipeline, time) would be interesting 🤔

  • @phizc
    @phizc Рік тому

    Workflow:
    1. Stable Diffusion XL (Watermark, 1024x1024, 400 images per day)

    • @ArtificallyIntelligentAi
      @ArtificallyIntelligentAi  Рік тому

      If anyone ask.. You didn't learn that here lol. I mean technically it's their own tool that did it so...ya.

    • @phizc
      @phizc Рік тому

      @@ArtificallyIntelligentAi I didn't, really. I paused at 6:40 to write my comment, so I hadn't seen you do it at the time 😄. Like minds think alike apparently 😅.

  • @Kryptonic83
    @Kryptonic83 Рік тому

    69th like, n i c e - Looking forward to trying out SDXL 1.0 in Auto1111 soon.

  • @GhostNameless
    @GhostNameless Рік тому +1

    Poor fairy stuck in a bottle. How dare you?!

  • @tradehut2782
    @tradehut2782 Рік тому

    While these images look great, they are not as mindblowing as what Leonardo's Alchemy gives.
    Alchemy is next level