Midjourney TRICK: Multi-Prompt Theme Infusion // Prompt Secrets of a PRO

Поделиться
HTML-код
  • Опубликовано: 31 дек 2024

Комментарии • 53

  • @rajeshkannanmj
    @rajeshkannanmj Год назад +13

    Prompts used :
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::autumn rain theme::1 --ar 2:3 --V 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::winter theme::1 --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::winter theme::1 --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::full moon night theme::1 --ar 2:3 --V 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::horror movie theme::1 --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::angry expression::1.5 --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::robotic wolf::1.5 --ar 2:3 --v 4
    also works with:
    robotic theme::0.65
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography:: 80s movie theme::1 --ar 2:3 --v 4
    paris alley with the eifel tower in the background::movie theme::1 --ar 2:3 --v 4

    paris alley::1 eifel tower in the background::2 cyberpunk theme::2
    a cat wearing a witch hat and robe::movie theme::1 --ar 2:3 --v 4
    a cat wearing a witch hat and robe::laboratory theme::0.65 --ar 2:3 --v 4
    a cat wearing a witch hat and robe::laboratory theme::0.65 --ar 2:3 --v 4

  • @RedRojo210
    @RedRojo210 Год назад +7

    the way you explained multi weighted prompts along with visual examples just made it click for me, thank you

    • @OlivioSarikas
      @OlivioSarikas  Год назад +3

      Thank you very much. Happy I could help you

    • @Mystsong18
      @Mystsong18 Год назад +1

      Me too! I couldn't figure it out before but now it makes sense. Thank you!

  • @CryptobdSchool
    @CryptobdSchool Год назад +1

    I think this year we will see lots of amazing AI tools that will make our lives easier. Those who can use these AI tools smartly will be the gainers. Recently I created amazing images with Blue Willow

  • @nahiddotai
    @nahiddotai Год назад +1

    I was watching one of Tokenized videos, and I remember him mentioning that without any weights, the default of each word in the prompt is basically 1. So, I'm surprised that by adding ::1 is actually adding more weight to that word. Every time I've used weights, I've always gone with 2 or higher but now I will try more images with weight 1. Thanks for sharing this awesome tutorial!

  • @_The_Wu_
    @_The_Wu_ Год назад +5

    MJ seems to have made a huge leap forward with realistic/detailed images using basic prompts lately. Guess it will get better and better over time.
    Been wondering how much MJ learns from prompts with uploaded images.

    • @Emsyaz
      @Emsyaz Год назад

      MJ is undeniably way better than SD

  • @eurossocial
    @eurossocial Год назад +2

    Excellent tutorial once more! Thank you for sharing Prompt tips... But I am wondering why people still keep writing the (--v 4)? It is already set in Midjourney by Default. It will not change the version unless you choose a lower versions below v4. Now Midjourney's version 5 has arrived. So for what reason people are still writing this parameter?

  • @balrajvishnu
    @balrajvishnu Год назад

    Learnt a few more techniques with Midjourney, thanks

  • @parrottm76262
    @parrottm76262 Год назад +1

    All the variations were totally mind blowing. Love the cat ones. 😃

  • @williamarenas11
    @williamarenas11 Год назад +1

    Thank you, as always you make amazing video’s.

  • @tinygenius
    @tinygenius Год назад

    Fantastic tutorial. I love it. Now it's time to experiment. 😃

  • @JessicaQChen-lv5uk
    @JessicaQChen-lv5uk Год назад +1

    Thank you. I really love your work!

  • @TheCultivatedBeat
    @TheCultivatedBeat Год назад +1

    Dude. How are you so helpful?!

  • @OlivioSarikas
    @OlivioSarikas  Год назад

    #### Win a RTX 3080Ti ####
    Sign up FREE to Nivida GTC for a Chance to Win a RTX 3080Ti:
    nvda.ws/3D7nNY4
    #### Links from the Video ####
    Follow www.youtube.com/@TokenizedAI
    Join on Facebook: facebook.com/groups/theairevolution
    Join on Discord: discord.gg/XKAk7GUzAW
    Robomar Hyper-Realistic Prompt Guide: ruclips.net/video/1Zked4DLPiQ/видео.html
    Prompts used: (Thank you to @rajeshkannanmj for transcribing them from the video)
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::autumn rain theme::1 --ar 2:3 --V 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::winter theme::1 --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::winter theme::1 --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::full moon night theme::1 --ar 2:3 --V 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::horror movie theme::1 --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::angry expression::1.5 --ar 2:3 --v 4
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography::robotic wolf::1.5 --ar 2:3 --v 4
    also works with:
    robotic theme::0.65
    wolf in a forest, professional color grading, soft shadows, no contrast, clean sharp focus, film photography:: 80s movie theme::1 --ar 2:3 --v 4
    paris alley with the eifel tower in the background::movie theme::1 --ar 2:3 --v 4

    paris alley::1 eifel tower in the background::2 cyberpunk theme::2
    a cat wearing a witch hat and robe::movie theme::1 --ar 2:3 --v 4
    a cat wearing a witch hat and robe::laboratory theme::0.65 --ar 2:3 --v 4
    a cat wearing a witch hat and robe::laboratory theme::0.65 --ar 2:3 --v 4

  • @wildcardnt
    @wildcardnt Год назад

    Really really liking your content a lot.

  • @tpeasetiger
    @tpeasetiger Год назад

    From my understanding just doing the :: applies a default weight of 1 anyway, so you don’t technically need to write the weight 1 for a weight of 1. So everything here with the :: after it has a weight of 1.

  • @NickSpirov
    @NickSpirov Год назад

    This is very interesting, and also there is a detail here - as you are misspelling Eiffel, it would be interesting to play with misspelling and see what (if any) significance it has. I have had very interesting results from accidental misspelling like this.

  • @eveFlower101
    @eveFlower101 Год назад +2

    I’ve been using the theme but not with weights. Going to test it out ❤

  • @dougrobinson2024
    @dougrobinson2024 Год назад

    @Olivio Sarikas - Are you just making whatever themes you want? Or are there specific ones that you need to use because Midjourney knows and understands them already?

  • @emmanuelgonzalezcaseira9141
    @emmanuelgonzalezcaseira9141 Год назад +3

    By all means, I'm interested in seeing how this would work with Stable Diffusion :), I'm still pretty new so having a reference would be great!

    • @roybatty2268
      @roybatty2268 Год назад

      Basically this multiprompt idea is like having a base model and then the second theme is a LORA trained on the theme. The problem is finding a LORA trained on the theme you are interested in. If you want Genshin Impact then you're set, but if you want autumn leaves you are going to have a harder time finding something already unless you make it yourself. There are luckily more how to make a LORA tutorials than a month ago.

    • @OlivioSarikas
      @OlivioSarikas  Год назад +1

      Loras is a good idea, but SD also has prompt weight, so you can change a summer to a winter scene with that without needing a lora. Loras are more useful if you need a specifically trained style, object or person

  • @poweron1115
    @poweron1115 Год назад

    thank you teacher❤

  • @davidpurple3698
    @davidpurple3698 Год назад

    Great video. Keep up the great work. What about how to install Stable Diffusion on a Mac the easiest way and how to use it ?

  • @himik511
    @himik511 Год назад

    dude you're really cool

  • @antonionieves4638
    @antonionieves4638 Год назад

    I love your tutorials so much. I will love if you can help me with something i am struggling, and it is, how to make images 3000 x 3000 with 300 DPI... i truly hope you can.

  • @yeremiahangeles7252
    @yeremiahangeles7252 Год назад

    How did you get good quality photos every time? And do you have to pay and how much if you don't mind me asking?

    • @OlivioSarikas
      @OlivioSarikas  Год назад

      I have the Standard Subscription for MJ. All the prompts are in the video and I added them to the Video Info to (thanks to @rajeshkannanmj for typing them down from my video)

  • @relaxingmusicgi
    @relaxingmusicgi Год назад

    You are the best

  • @musicpet
    @musicpet Год назад

    Show this in SD plz 😊

  • @yeremiahangeles7252
    @yeremiahangeles7252 Год назад

    What does the numbers mean and do?

    • @OlivioSarikas
      @OlivioSarikas  Год назад +1

      The number after the :: is the weight. It means how much importance something in the image has. So if you typ dog::2 in a park::1 the dog should be bigger and if you type dog:1 in a park::2 the park should take more space in the image

  • @joshualibretto
    @joshualibretto Год назад

    Wow

  • @evelynannrose
    @evelynannrose Год назад +1

    tokenized is a G

  • @BigMikeMora
    @BigMikeMora Год назад

    This is why I need MJ....SD just doesn't do it all ...

  • @Pauluz_The_Web_Gnome
    @Pauluz_The_Web_Gnome Год назад

    Too bad MJ is too expensive! I'll stick to A1111

    • @OlivioSarikas
      @OlivioSarikas  Год назад

      Yes, true, but at the same time it is easier to use

  • @SALLI.E
    @SALLI.E Год назад

    Oh my God... I absolutely love your videos... you helped me very much... your faithful first AI fan @SALLI.E