New AI - Image to 3D

Поделиться
HTML-код
  • Опубликовано: 1 фев 2025

Комментарии • 61

  • @jameschen7822
    @jameschen7822 11 дней назад +44

    another greate opensource project from China

    • @sebastiankamph
      @sebastiankamph  11 дней назад +5

      For sure, they've been releasing some great stuff!

    • @yikifooler
      @yikifooler 11 дней назад +3

      not only releasing but releasing plenty of amazing ones with opensource like Imagine they could earn money but they are not doing that just giving It free away.

    • @juanjesusligero391
      @juanjesusligero391 5 дней назад +1

      It's a nice model and all, but it's not open source.
      Unlike other licenses like Apache or MIT, the "Hunyuan Community License" imposes significant restrictions. For example, if you are in Europe, the license does not apply to you (which, as I understand, means you cannot legally use the model).

    • @yikifooler
      @yikifooler 5 дней назад

      ​@ European people are jamming Hunyuan on their local PC and you are complaining about restrictions, kids these days uses Dark Web on their fathers pc without anyone knowing and sleeps tight at night and yet you are complaining something which is not legal, London is allowing Muslims to rule their Kingdom that should be illegal privately too.

    • @dollynhooetdevarginha8
      @dollynhooetdevarginha8 5 дней назад

      @@juanjesusligero391 where can we use the license? is brazil allowed? lol

  • @Elwaves2925
    @Elwaves2925 11 дней назад +3

    I stumbled on the demo page just before you posted, tried a Lego Minifig image and it did a great job. Not perfect as the legs were screwed up but it worked really well for the hair considering there was only a front view.
    Just got to see how well it does textures but it's looking very promising.

    • @sebastiankamph
      @sebastiankamph  11 дней назад +1

      Very promising indeed! Hunyuan has been releasing some amazing stuff 💫

  • @lucasmarques.mp4
    @lucasmarques.mp4 4 дня назад +1

    I just tried it, and in order for the text prompt to work, you have to remove the image... It doesn't care which tab you selected, if you have an image in the first tab, it will take that as the prompt!

  • @nartz5361
    @nartz5361 10 дней назад +12

    the demo now is 404 error page 😭😭😭😭

  • @caglayangursugokbulut9664
    @caglayangursugokbulut9664 4 дня назад

    Great tutorial. Can we only use the demo version at the moment?

  • @Kinos141
    @Kinos141 11 дней назад +3

    Raphael was my favorite ninja turtle.
    Also, this is really cool.

    • @sebastiankamph
      @sebastiankamph  11 дней назад

      Does that mean we're mortal enemies now?

    • @mik3lang3lo
      @mik3lang3lo 10 дней назад

      I don’t trust anyone who doesn’t say Michelangelo

    • @kboid5919
      @kboid5919 11 часов назад

      @@mik3lang3lo Midhelangelo was a clown. Only Leonardo.

  • @chariots8x230
    @chariots8x230 10 дней назад +3

    7:00 It would be useful if there was some editing capabilities within the program, so you can select the part that you want to edit, such as the turtle’s toes, and fix the mistake by prompting the AI to give the turtle the correct number of toes.

    • @AscendantStoic
      @AscendantStoic 10 дней назад +2

      Suppose that's what Blender is for 😅
      Nvidia also released an AI recently that vastly improves the polygons of 3d models and optimizes them.

  • @importon
    @importon 5 дней назад +6

    this video should have been 2 mins

  • @niclasm4123
    @niclasm4123 11 дней назад +3

    1 minut sedan! Go Sebastian! Dad Jokes 😂

    • @sebastiankamph
      @sebastiankamph  11 дней назад +2

      Vroooooooooom! I promise, there's at least one dad joke.

  • @j.e.711
    @j.e.711 4 дня назад

    I tried to generate a model of a knife handle with scrollworks and the result of this software compared to SculptOK is terrible bad, Maybe it works well for the images on the right side but not for uploaded images.

  • @AndreyJulpa
    @AndreyJulpa 10 дней назад +9

    Waiting for comfy

  • @lianmdw
    @lianmdw 7 дней назад

    Have you seem the new 3d model called Spell from spline?

  • @Lv7-L30N
    @Lv7-L30N 10 дней назад

    Thank you

  • @DarnIDidntKnowThat
    @DarnIDidntKnowThat 4 дня назад

    it doesnt work on local pc its a head ache to install

  • @pedroserapio8075
    @pedroserapio8075 10 дней назад

    Interesting, thank you. Also getting random 500 error.

  • @g0d182
    @g0d182 День назад

    nice

  • @leandrogoethals6599
    @leandrogoethals6599 10 дней назад

    The demo has runtime error

  • @seanknowles7987
    @seanknowles7987 10 дней назад

    maybe i missed ti but is it opensource? Is there an API for this?

  • @lowaura
    @lowaura 4 дня назад

    Does this work locally?

  • @Wreighn
    @Wreighn 4 дня назад

    Does it work on amd?

  • @jonmichaelgalindo
    @jonmichaelgalindo 10 дней назад

    Huh. In my first demo test (own pic) it was WAY worse than Trellis. Disgustingly bad. Maybe the demo's default settings were bad? Hmm. I'll try again when it's in Comfy.

  • @TelefonBrick
    @TelefonBrick 4 дня назад

    🐋 hello there!

  • @Sketching4Sanity
    @Sketching4Sanity 11 дней назад

    LOVE

  • @NITHZAEL92
    @NITHZAEL92 10 дней назад

    Error 500 😢 on demo page

  • @chariots8x230
    @chariots8x230 10 дней назад +2

    Krea’s image to 3D, and back to image is also very cool. It came out recently and it’s worth checking out.

    • @sebastiankamph
      @sebastiankamph  4 дня назад

      As far as I know they only use open source solutions packaged into a user interface.

  • @SCHOOLERstyle
    @SCHOOLERstyle 9 дней назад +4

    💪i create better 3D models on my channel than any AI generated JUNK 💪

    • @TheBikeOnTheMoon
      @TheBikeOnTheMoon 8 дней назад +2

      kinda alarming that these AIs is getting into 3D stuff now. My friend shows me that AI can even do PBR texture eventhough it kinda looks like shit atm, but gives it a few more years and 3D artists would be out of business. The future is dark for all artists.

    • @lespectator4962
      @lespectator4962 5 дней назад +2

      ​@@TheBikeOnTheMoonIP laws will just adjust to this. This tech still needs human input to improve. Idk why people are so excited to kill off artists because this tech will stagnate without them.

  • @petermclennan6781
    @petermclennan6781 4 дня назад

    Interesting, but limited usefulness, IMHO. The mesh is all one file, right? No components. Impossible to rig for animation.

  • @iAmAlbert
    @iAmAlbert 10 дней назад

    404 on the demo?

  • @electronicmusicartcollective
    @electronicmusicartcollective 8 дней назад

    Lieber Sebastian, bitte poste erst ein Video wenn Du es in ComfyUI erstellen kannst. Videos wie diese gibt es schon viele, auch KI generiert. Alles gute

    • @sebastiankamph
      @sebastiankamph  4 дня назад

      You can. It was available shortly after. Link in description.

  • @360sblulev
    @360sblulev 10 дней назад +1

    cant wait for hunyuan img2vid

    • @photobackflip
      @photobackflip 10 дней назад

      FYI - there are a few img2vid workflows on Civitai.

  • @Drflow_Mwemsyy
    @Drflow_Mwemsyy 4 дня назад

    He was key what? 😹

  • @praiseandworship_com
    @praiseandworship_com 10 дней назад +1

    $$$$ HELP!!!! Anyone out there want to help me with a 6 minute music video? Paid gig. good artist!!

  • @khanra17
    @khanra17 4 дня назад +1

    You are intentionally trying hard to be annoying 😡

  • @BrostopherWade
    @BrostopherWade 9 дней назад

    Kijai has updated ComfyUI-Hunyuan3dWrapper