[0:00] hey today I want to show you how I
[0:01] prompt in stable diffusion how I think
[0:03] about it when prompting and what methods
[0:05] I use to create prompts faster I'm using
[0:08] stable diffusion Forge UI and Juggernaut
[0:11] XL version 10 model but you can use any
[0:13] model you prefer with the settings that
[0:15] work best for that model many beginners
[0:18] simply prompt something simple like
[0:19] portrait of a woman while many short
[0:22] prompts still work in interfaces like
[0:24] Focus or AI like mid journey and chat
[0:26] GPT you're leaving too much freedom to
[0:29] the AI
[0:30] you can be more specific to get
[0:32] something closer to what you have in
[0:33] mind for example you can specify the
[0:35] type of image you want such as a photo
[0:37] illustration painting and so on as you
[0:40] can see even by mentioning photo the
[0:42] possibilities are still too many I got a
[0:44] vintage photo so maybe you can add an
[0:47] extra word like modern to specify that
[0:49] you want a modern photo instead of a
[0:52] vintage one to experiment better with
[0:54] the prompt you can use a fixed seed you
[0:57] can either use the last seed or just
[0:59] type number and generate again so we
[1:03] have the type of image and a subject but
[1:06] let's also place the subject in an
[1:08] environment maybe we can put her in the
[1:10] forest or on a beach or perhaps we want
[1:12] a studio photo with a black background
[1:14] behind maybe I don't want the woman to
[1:17] have a random hair color so I add blonde
[1:20] hair in the prompt perhaps I want her
[1:22] shirt to always be white so it stands
[1:24] out against the black background maybe I
[1:26] want the image to pop out more so I can
[1:29] add a rim light coming from behind
[1:32] visible in the hair alternatively I
[1:34] might want a golden hour type of light
[1:36] to make it look more natural if you have
[1:39] a specific hairstyle in mind you can add
[1:41] that to the prompt if you don't know the
[1:43] name you can search on Google for
[1:45] hairstyles women names for example if
[1:48] you like bangs you can add that to the
[1:51] prompt Additionally you can use chat GPT
[1:54] to provide lists for anything such as a
[1:56] list of women's clothing so I can use
[1:58] like a red cardigan instead of a white
[2:00] shirt let's try different types of art
[2:03] styles for the photo instead of just
[2:05] saying painting portrait which is a
[2:07] broad category we can specify further
[2:10] for example it could be an oil painting
[2:12] maybe with an old look or perhaps it's a
[2:15] watercolor painting we could also
[2:17] consider a pencil drawing but in this
[2:19] case we'd need to adapt the colors and
[2:21] background to match that style uh let's
[2:24] play around more with photography a
[2:27] cinematic fashion portrait could look
[2:29] cool
[2:30] we can also mention the nationality of
[2:33] the woman or the country for instance we
[2:36] could generate in African woman or a
[2:39] police woman since we've been so
[2:42] specific with clothes it might not
[2:44] include the police uniform so instead of
[2:46] a cardigan let's add a police suit or
[2:49] uniform let me quickly show you the
[2:52] negative prompt here you list things
[2:55] that you don't want to appear in your
[2:57] image it doesn't work perfectly some
[3:00] words are stronger than others but for
[3:02] example if you don't want a police badge
[3:05] you can add that to the negative prompt
[3:07] now when I generate again the badge is
[3:09] gone if you don't like the blue light
[3:11] behind you can add it to the negative
[3:13] prompt if that doesn't work you can try
[3:15] different seeds um in the worst case
[3:18] scenario you can use in painting to
[3:20] remove it let me remove the negative
[3:22] prompt and I'll show you how you can
[3:24] replace different words in the prompt
[3:25] scroll down and go to the XYZ plot on
[3:28] the X Type search for prompt Sr on the X
[3:31] values you put the words you want to
[3:33] search and replace first so for example
[3:36] I want to search for the word blonde and
[3:38] replace it with different colors like
[3:40] dark purple red and so on this way you
[3:43] can see how the girl looks with
[3:45] different hair colors as you can see now
[3:47] it started to generate the policewoman
[3:49] with different hair colors as a bonus at
[3:52] the end you have a grid with the
[3:53] searched word and all the other
[3:55] variation words that were replaced you
[3:57] can open it and see how they look in
[3:59] different
[4:00] colors let me dress her in a white shirt
[4:02] real quick let's say I want to redo
[4:04] everything but the girl's face to be
[4:06] more similar between Generations one
[4:09] thing that helps is to give her a name I
[4:11] usually use a random invented name but
[4:14] you can also use a mix of celebrities if
[4:16] you want so I'll give you a name and
[4:19] I'll generate again with that search and
[4:21] replace script active as you can see in
[4:24] the results the woman is quite similar
[4:26] between
[4:27] Generations because it has the same seed
[4:30] and the same description and name only
[4:33] the color is different let's say you
[4:35] like the results of a prompt but you
[4:37] just want a very subtle variation of it
[4:40] you can do that by adjusting the
[4:42] sampling steps or CFG scale you can do
[4:45] that automatically if you go back to the
[4:46] scripts and instead of prompt search and
[4:49] replace you choose CFG scale for the X
[4:52] values you can put between five and
[4:54] seven for example now when you generate
[4:57] you'll get those very subtle variations
[4:59] they might be harder to see on the
[5:00] screen but let's go to the folder where
[5:02] they were saved and switch from one to
[5:04] another as you can see some don't have
[5:06] badges or the light is a little
[5:08] different so you can choose the best one
[5:10] from all or you can combine them in
[5:12] Photoshop and keep what is best from
[5:15] each I also like to get variations for
[5:18] prompts using chat GPT so I have this
[5:20] prompt that I know works okay and I want
[5:23] a prompt adapted from different
[5:25] jobs I can ask GPT to adapt it for me
[5:29] and as you can see I get some nice
[5:32] variations let me test quickly the
[5:34] doctor prompt as you can see it added a
[5:37] white coat to the prompt now let's see
[5:39] how it looks like as a chef copy and
[5:41] paste The Prompt using the control plus
[5:43] C and Control Plus V shortcuts and
[5:46] generate quickly different variations
[5:49] sometimes I'm too lazy to write
[5:51] descriptive prompts so I just ask chat
[5:54] GPT to write it for me you just need to
[5:56] guide it a little in the right direction
[5:59] let's test it really quick to see what
[6:00] we get seems to work okay go to image to
[6:03] image and let me show you another trick
[6:06] that you can use when you don't know how
[6:07] to prompt but you have a photo or an
[6:09] illustration upload it here and then
[6:11] click on the paperclip icon that will
[6:14] interrogate clip which stands for
[6:16] contrastive language image pre-training
[6:18] the first time you're using it it will
[6:20] take some time to download a model you
[6:22] can see when it's ready in the command
[6:24] window but after that it should
[6:26] interrogate quicker that will give you a
[6:29] prompt for that image I'll copy this
[6:31] prompt and test it quickly the result is
[6:34] not bad could be better you can see that
[6:37] it's a policewoman but she doesn't look
[6:38] very similar to my image what seems to
[6:41] work better and get more consistent
[6:44] results for me is chat GPT I just upload
[6:47] a photo to chat GPT and ask it to
[6:49] describe it for me in a long sentence
[6:52] let's copy The Prompt and test it to see
[6:53] what we get now look at that how similar
[6:56] the photos look like with the image I
[6:58] gave it
[7:00] now you can talk with chat GPT to change
[7:02] the prompt how you want it for example
[7:04] you can adapt it for a doctor instead
[7:07] and this is the result if you don't have
[7:09] access to GPT or the internet and you
[7:12] still have problems with uh prompting
[7:14] but want to get some nice generation you
[7:17] can use art styles instead so if you
[7:20] have a short prompt you can go to art
[7:22] styles and either save your own art
[7:25] styles or download some free art styles
[7:27] from the internet I have an entire video
[7:29] with with 260 art styles and how to
[7:31] download them if you look at the prompt
[7:33] description then you look at the info
[7:35] for Generation you can see that you get
[7:38] the original prompt first then the art
[7:40] style is just another already saved
[7:42] prompt that is added to the original
[7:44] prompt plus the negative prompt let's
[7:47] talk a little bit about how to add more
[7:49] weight to certain words I have this
[7:52] prompt where I added green blue house
[7:54] next to each other to show you how I can
[7:56] make one word more important than the
[7:57] other so this is what I get for this
[8:00] prompt the house has blue and green in
[8:02] it I can make the words Blue House more
[8:04] important if I add round brackets you
[8:07] can see if you forgot to close a bracket
[8:10] it shows in red in the corner where the
[8:12] tokens are once I generate you can see
[8:15] the house has more blue than green you
[8:17] can add even more weight to those words
[8:19] if you add another
[8:21] bracket but there's an easier way to do
[8:24] this faster you just select the words
[8:27] then press control and use the up Arrow
[8:29] key to add more
[8:30] weight with control and the down arrow
[8:33] you can reduce the weight it can also
[8:36] have a negative value if you want it to
[8:37] be less important to change the value
[8:40] you select only the text without the
[8:42] brackets and numbers and use the
[8:44] shortcut again alternatively you can
[8:47] just change the numbers manually you
[8:50] should avoid going too high usually if
[8:52] you go more than 1.9 or two values it
[8:56] starts to go a little crazy and you get
[8:58] something like this like an a I exploded
[9:00] brain if I add a negative value to The
[9:02] Gnome for example it will be less
[9:05] important and you can see in the results
[9:07] that it's barely visible the main
[9:10] subject becomes the
[9:12] house here's how I usually prompt I put
[9:15] the art style or medium either first or
[9:17] last then comes the subject followed by
[9:19] the description then the environment and
[9:22] finally any extra information like
[9:23] colors lighting and mood while I was
[9:26] working on the video chat GPT released a
[9:29] new model version GPT 40 so now it's
[9:32] even easier to just ask for a stable
[9:35] diffusion prompt for what you want to
[9:36] create so if I want a watercolor
[9:38] painting of a cute bunny with glasses I
[9:40] can get a prompt quickly now I can test
[9:43] it in stable diffusion and the result is
[9:45] what I asked for I can then make more
[9:48] changes for example I can ask chat GPT
[9:51] to make it more minimalist and add a
[9:53] white background so now when I test it I
[9:55] get a nice cute cartoon bunny on a white
[9:58] background another thing you can do is
[10:00] to turn on generate Forever by right
[10:02] clicking on the generate button and
[10:04] choosing generate
[10:06] forever to stop it you need to
[10:08] rightclick again and choose cancel
[10:10] generate forever if you don't want to
[10:12] let it generate forever and have a
[10:14] specific number of generations in mind
[10:16] like between one and 100 you can move
[10:19] the batch slider to the value you want
[10:22] and it will generate that number of
[10:24] images alternatively if you have
[10:27] multiple different prompts you can
[10:29] choose prompts from a file or a text box
[10:32] you can paste prompts here or upload a
[10:34] text file with the prompts just make
[10:35] sure each prompt is on a separate line
[10:38] um I can once again use chat GPT to give
[10:40] me a list of different variations of the
[10:42] prompt with different animals then I can
[10:45] uh copy and paste that list of prompts
[10:47] in the text area when I hit generate it
[10:50] will start generating each of those
[10:52] prompts um as you can see I get
[10:53] different animals on a white background
[10:56] if you have any questions you can post
[10:57] them in my Facebook group pix Roma
[10:59] Community I usually post news prompts
[11:02] daily challenges design and crafts there
[11:05] I've recently passed 1,000 members so
[11:07] thank you to all who joined if you found
[11:09] something useful in this video leave a
[11:11] like thank you and have a great day