Comments

Blacklisted
    (help)
    Show Search Options
    post #53738
    156 269 1 E
    post #11898
    72 148 1 E
    post #9730
    6 10 5 S
    post #53564
    126 174 4 Q

    user 20054

    Member

    oaf40 said:
    I recommend you to give Canny & Depth modules a try. They make working with "real" backgrounds much easier in my experience, and the restyling capabilities are very flexible. Here's my workflow just as an example (apologies for typos as English isn't my native language):

    (before upscaling)

    1. Perhaps you found a real-life photograph you'd like to use as background but you don't like how it would look with your current styling
    2. Load the background photo into img2img, select your model. The prompt (including the negative one) should be a brief description of what's on the photo. In my experience 50-75 tokens in the positive prompt is enough. If your model uses score tags then use them here as well. Don't forget to include all your style LoRAs!
    3. If the photo is bigger than the working resolution of the model you use then downscale it using the "Resize to/by" submenu or in an external editor, this is important
    4. Sampler/steps: I use either DPM++ 2M @ 40 steps or Euler A @ 80 steps for this purpose
    5. Activate the ControlNet plugin, you'll need 2 units. First one would be Canny, second one would be Depth. As for the module files I recommend using models by Xinxir as they work the best with PDXL-based base checkpoints and mixes. You can easily find them on Hugging Face.
    6. For each unit tick the "Upload independent control image", upload the same source photo. Select the preprocessor ("canny" for Canny, "depth_midas" for Depth), make the resolution higher (bumping 512 up to 960 works fine with 16 GB of VRAM), set Control Weight to 1. Set "Ending control step" to 0.3 (Unit will self-deactivate at 30% of steps) if you use plain styles or somewhere around 0.7 if you use "realistic models". As for the two Threshold sliders for Canny I recommend trying various values to see which would preserve enough details for you, they vary for each picture you're working with. Click the smol button with the "Bang" emoji on it, save both Canny and Depth masks, set the Preprocessor to "None" and reupload freshly produced masks again - this would save you from running preprocessors every generation
    7. Play around with denoise strength, in my personal experience 0.7~0.83 works the best for plain styles.
    8. Click "Generate" (if you are on A1111 ofc) and it will get you a restyled photo that would fit your overall style.

    After you combine the picture of your character(s) with the background repeat the same procedure with following considerations:
    1. Merge the prompts for both the background and your characters, use BREAK to separate keywords
    2. Use much lower denoising strength in img2img, 0.1~0.2 should be enough to blend the background and foreground and kill some artifacts and jagged lines without affecting the details
    When done, proceed with your usual upscaling and post-upscaling routines... :P Maybe there are better alternative I don't know about yet (I'm sure as hell there are) but for me Canny&Depth were The life changers.

    Ohh Thanks! I've read about this technique recently but didn't try this yet. I appreciate your help a lot, will give it a try!

  • |
  • 0
    •
    post #53712
    156 287 2 E
    post #17937
    102 158 1 E
    post #53159
    117 192 3 E

    September

    Member

    jelloponies said:
    Hiya!
    I sometimes post stuff I don't upload here on the furrydiffusion discord but that's about it.

    Cool! Maybe you know how to get access there? Invite or something. Your works are just too great, so I would like to see everything <3

  • |
  • 0
    •
    post #47998
    89 120 1 Q
    post #52409
    41 53 6 E

    Emeraldwolf

    Member

    YetAnotherAIuser said:

    emeraldwolf said:
    I tried using your tags on a different ai art thing but some return like a weird thing what ai did you use?

    Based on offline Stable Diffusion web UI

    According to the prompt info you need also to install
    Checkpoint AI model: ponyDiffusionV6XL_v6StartWithThisOne
    LoRA: Photo 2 Style SDXL_LoRA_Pony Diffusion V6 XL

    Link?

  • |
  • 0
    •
    post #53485
    126 177 3 E
    post #53704
    87 135 4 E

    Foxymutt

    Member

    birdlover said:
    Already done. I overdid it a little though. I guess I have around 10-15 pics which I think are postworthy. Don't want to spam e6 though. Need your help in selecting the good ones (I'm a not a dog expert :> )

    I'll see if I can upload them somewhere and send you a link Tomorrow

    Space them out over e6 here and there and you decide what you think the best just looking forward to what you have in mind 😊

  • |
  • 0
    •
    post #53704
    87 135 4 E

    BirdLover

    Member

    foxymutt said:
    Australian shepherd female plz 🙏

    Already done. I overdid it a little though. I guess I have around 10-15 pics which I think are postworthy. Don't want to spam e6 though. Need your help in selecting the good ones (I'm a not a dog expert :> )

    I'll see if I can upload them somewhere and send you a link Tomorrow

  • |
  • 0
    •
    post #53704
    87 135 4 E
    post #52990
    70 99 3 E
    post #53703
    118 172 2 E
    post #53562
    113 144 4 E

    Pervo

    Blocked

    this is your first? you don't get any better and higher quality than this in 2024, this is excellent

  • |
  • 0
    •
    post #53705
    57 75 1 E
    post #53568
    169 270 4 E
    post #53568
    169 270 4 E
    post #53616
    78 96 2 E

    HorseFucker5

    Member

    Even if she does have paws, this is one of the cutest ponies Ive seen in a long time! I really want to see more of her!🦄🤩👌

  • |
  • 2
    •
    post #53691
    79 105 3 E

    Foxymutt

    Member

    blp said:
    parking airships is harder than it looks. :( best leave it to the professionals.

    Well I tryed she don't look to mad tho lol

  • |
  • 1
    •
    post #52409
    41 53 6 E

    machine yearning

    Member

    yetanotheraiuser said:
    Based on offline Stable Diffusion web UI

    According to the prompt info you need also to install
    Checkpoint AI model: ponyDiffusionV6XL_v6StartWithThisOne
    LoRA: Photo 2 Style SDXL_LoRA_Pony Diffusion V6 XL

    This was entirely inpainting (and some gimp editing). The description prompt and the image prompt are just from the final inpainting step, but the bulk of it remained unchanged throughout except to accumulate new features as they were added.

  • |
  • 0
    •
    post #53676
    100 136 1 Q
    post #53691
    79 105 3 E

    blp

    Member

    foxymutt said:
    Love her expression she's like last time I let you park the airship 😸

    parking airships is harder than it looks. :( best leave it to the professionals.

  • |
  • 1
    •
    post #53691
    79 105 3 E
    post #52397
    218 358 5 Q
    post #52812
    294 493 4 E
    post #53670
    107 153 1 E
    post #45439
    186 305 1 E
    post #53461
    42 85 3 E
    post #53608
    78 112 1 E
    post #53623
    42 63 2 E
    post #53533
    46 66 1 E
    post #53562
    113 144 4 E
    post #53400
    59 87 2 E
    post #53684
    68 118 2 E

    Fenix AI

    Member

    jons said:
    Multiple resulted in a pretty result

    Indeed, to be fair only the blue eye was original, the other one was edited from another picture that I preferred the look of as part of the same generation, liked the way it turned out so i put it on this picture

  • |
  • 2
    •
    post #53256
    167 269 2 E
    post #53684
    68 118 2 E
    7 7 4 Q
    17 34 7 E
    post #51152
    109 173 2 E
    post #53497
    246 410 3 E
    post #50970
    93 126 7 E

    blp

    Member

    woof said:
    if she is my navigator, I'm in!

    don't forget the DOG IS MY NAVIGATOR bumper sticker!

  • |
  • 0
    •
    post #52994
    121 214 1 E

    Zaddha

    Member

    I don't know who posted my messages, but I'm grateful to you. I'm too lazy to tag art

  • |
  • 1
    •
    post #53564
    126 174 4 Q

    oaf40

    Member

    k3y said:
    It worked, but for whatever reason it became slightly blurry... gotta work on that some more.

    I recommend you to give Canny & Depth modules a try. They make working with "real" backgrounds much easier in my experience, and the restyling capabilities are very flexible. Here's my workflow just as an example (apologies for typos as English isn't my native language):

    (before upscaling)

    1. Perhaps you found a real-life photograph you'd like to use as background but you don't like how it would look with your current styling
    2. Load the background photo into img2img, select your model. The prompt (including the negative one) should be a brief description of what's on the photo. In my experience 50-75 tokens in the positive prompt is enough. If your model uses score tags then use them here as well. Don't forget to include all your style LoRAs!
    3. If the photo is bigger than the working resolution of the model you use then downscale it using the "Resize to/by" submenu or in an external editor, this is important
    4. Sampler/steps: I use either DPM++ 2M @ 40 steps or Euler A @ 80 steps for this purpose
    5. Activate the ControlNet plugin, you'll need 2 units. First one would be Canny, second one would be Depth. As for the module files I recommend using models by Xinxir as they work the best with PDXL-based base checkpoints and mixes. You can easily find them on Hugging Face.
    6. For each unit tick the "Upload independent control image", upload the same source photo. Select the preprocessor ("canny" for Canny, "depth_midas" for Depth), make the resolution higher (bumping 512 up to 960 works fine with 16 GB of VRAM), set Control Weight to 1. Set "Ending control step" to 0.3 (Unit will self-deactivate at 30% of steps) if you use plain styles or somewhere around 0.7 if you use "realistic models". As for the two Threshold sliders for Canny I recommend trying various values to see which would preserve enough details for you, they vary for each picture you're working with. Click the smol button with the "Bang" emoji on it, save both Canny and Depth masks, set the Preprocessor to "None" and reupload freshly produced masks again - this would save you from running preprocessors every generation
    7. Play around with denoise strength, in my personal experience 0.7~0.83 works the best for plain styles.
    8. Click "Generate" (if you are on A1111 ofc) and it will get you a restyled photo that would fit your overall style.

    After you combine the picture of your character(s) with the background repeat the same procedure with following considerations:
    1. Merge the prompts for both the background and your characters, use BREAK to separate keywords
    2. Use much lower denoising strength in img2img, 0.1~0.2 should be enough to blend the background and foreground and kill some artifacts and jagged lines without affecting the details
    When done, proceed with your usual upscaling and post-upscaling routines... :P Maybe there are better alternative I don't know about yet (I'm sure as hell there are) but for me Canny&Depth were The life changers.

  • |
  • 1
    •
    post #53615
    37 47 2 Q

    ZeroOverZero

    Member

    The ruffled chest feathers sticking out adds this extra layer of detail that brings thoughts of the tactile sensation of feathers to my imagination. Or in other words, it really adds to the fantasy.

  • |
  • 3
    •
    post #53159
    117 192 3 E

    Jelloponies

    RATLOVER

    september said:
    Do you post your works somewhere else?

    Hiya!
    I sometimes post stuff I don't upload here on the furrydiffusion discord but that's about it.

  • |
  • 1
    •
    post #53601
    87 142 1 E
    post #53655
    157 267 3 E

    Shanti

    Blocked

    "Hey guys, did you know that in terms of male human and female Pokémon breeding, Vaporeon is the most compatible Pokémon for humans?"

    If such "jokers" would not repeat the same "joke not of the first freshness", few people would know about it.
    Therefore, such questions are as meaningless as possible. They are on the same shelf, with the joke "to poke a man's face in the pie

    I don't want to offend anyone, but repeating the same thing is not good.

  • |
  • -8
    •
    post #53549
    233 366 2 E
    post #53469
    80 151 2 E
    post #8390
    36 55 1 Q
    post #53499
    77 110 5 E

    shark enthusiast

    Member

    caserus said:
    Not into insects in general, but every once in a while one comes along and isn't nightmare fuel. Usually it's the bees that turn out really cute.

    In all fairness I left out some of the more traditional insect prompts making these more anthro/furry than insect, not a big fan of the more accurate insect stuff myself.

  • |
  • 0
    •
    post #52962
    231 385 3 E

    UnkindledFlames

    Member

    slops said:
    How many legs does that guy have again?

    4 legged friend, but you can only see 3.
    From the left, is the back rear leg, front left leg, front right leg

  • |
  • 0
    •
    post #53618
    56 85 1 E
    post #53562
    113 144 4 E
    post #53614
    97 131 3 E
    post #53575
    65 117 1 E
    post #53568
    169 270 4 E
    post #53644
    54 77 1 Q
    post #53152
    119 180 5 E
    post #53474
    64 86 2 E
    post #53639
    77 97 1 E
    post #53327
    58 97 2 E

    BirdLover

    Member

    dra-suzumebachi said:
    technically this should be tagged crossgender since only male peacocks have those vibrant colors and eye patterns on their tail feathers

    Ha, you're right. I'm fine with it, so if people agree, feel free to tag it as such :)

  • |
  • 0
    •
    post #53621
    57 74 2 E

    BirdLover

    Member

    mikubraver said:
    What about nimbats?

    Hmm. I'll try. Early results don't look good unfortunately. Prompting for a nimbat gives me rabbits instead. I'll try a few more checkpoints and hope for the best

  • |
  • 0
    •
    post #52723
    130 161 8 E
    post #53446
    44 54 2 E

    BirdLover

    Member

    Stunning!

    Great lighting and oh my god, that face, those eyes and that tail. Just pure perfection!

  • |
  • 1
    •
    post #53615
    37 47 2 Q
    post #53625
    97 112 5 E
    post #53625
    97 112 5 E
    post #53621
    57 74 2 E
    post #52374
    122 182 2 E
    post #53497
    246 410 3 E
    post #53568
    169 270 4 E
    post #53159
    117 192 3 E