5 Prompting Stable Diffusion Extension In 10 Minutes (Automatic1111)

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
if you're looking for extensions which can change the way you write prompts then stay tuned as I'll be exploring five prompt Focus extensions which could help improv the way you generate images I'm using version 1.6 of the web UI as I tend to use older versions in the event and extension has an updated to the latest version of the web UI this will be a brief overview of the extension rather than a full tutorial so let me know if there's any content you want covered in full but hit that like button and let me give it to you bite size so our first extension is SD Dynamic prompts and this extension allows you to test multiple prompts at the same time by adjusting your base prompt and utilizing wild cards sort of expanding on the features found within the prompt Matrix you can install this extension by navigating to the extensions tab within the web UI and searching for the extensions name under the available tab after loading the extension index once installed you will see two sections for the extension a wild cards manager on the top Riven and and an accordion below the seed within your user interface starting with the wild cards manager this tab allows you to edit wild card files either pre-installed or created yourself from within the web UI these wild cards can be called using a double underscore before and after the file name within your prompt section then you have the dynamic prompts Tab and this will have a few options for you to utilize the first is prompt magic which is a prompt generation model that modifies your prompt with random prompts which can change the style and quality of the image the next option is Ginger 2 templates which is how you can program templates for prompts using various functions like if statements variables and other functions similarly to a programming language but the meat of the extension is within the prompts itself so starting with a basic example you should first enable combinational generation within the extension so we can run test on a combination of prompts if I were to prompt in banana split with cherries then we'll get get that corresponding image but what if I wanted to see what this would look like with other fruits like grapes nuts or even syrup we can achieve this by modifying our prompts to use curly braces which would include the prompts we want to swap out and a pipe to separate the prompts being swapped now we have multiple images with our bass prompt Banana Split then the different types of toppings we specified appearing in each image you can also do more complex prompting using nesting which involves selectively swapping out our prompts within a prompt you're swapping out already or in other words adding additional information to our prompts going back to our previous example we ran a test on banana split with grapes nuts and syrup but if we wanted to test different types of nuts like cashew nuts almond nuts and even hazelnuts that we could change our prompt to look something like this this will now result in multiple images given us a banana split with grapes then cashew nuts almond nuts hazelnuts and then syrup the benefit is that we can add in those additional types of nuts to our existing prompt instead of having to create new prompts to add in that new information and to make things even better Dynamic prompts ignore whites space in TX meaning you can structure your prompt for easier reading as it grows making it easier to modify and understand later on you can go into even more depth with the combination of prompts you can use but as this is a high level overview I'll finish by saying that this extension also supports wild cards so you can have a tech file with words and then use wild cards to randomly select a word in that file to use as a prompt which can be useful for testing different types of prompts without having to manually adjust things each time wild cards are stored within the extensions Wild Card folder and are called by placing two underscores before and after the file's name within your prompt if you want me to do an in-depth review of this extension then let me know in the comment section below we can go through the different layouts use cases and techniques to make the most of this amazing tool our next extension is clone cleaner and this modifies prompts with random names nationalities and other factors automatically so you can get more variation in people generated without much manual work you can install the extension by entering the GitHub URL into the extensions installation Tab and then restarting the web UI the extension will be beneath the seed in an accordion and automatically enabled and you can modify a number of settings before you start generating your images for example you can specify whether to include names country hair length hairstyle and other elements in your image alongside using other settings to alter the results you get and when you generate an image you can see the prompts Us in the generation data meaning that when you find something you like you can reuse the information to generate further images this can also be useful for seeing how certain prompts like hairstyles affect the character's look and wherever those prompts are recognized by stable diffusion at all but this is a fairly simple extension that has the benefit of being both beginner friendly and a powerful tool if I generate a few images you can see we're getting variety without having to do much manual work and with an understanding of prompting techniques you can take what the extension provides and use it as a base to build on next we have tag autocomplete and this will provide you with autoc completion hints for recognized tags from image breu boards like damaru similar to suggestions your phone will give you when typ in text messages you can install this by searching for the extension in the list of available extensions within the web UI and it'll be under the name Buu Tag Auto completion the extension doesn't have any menu but it does have an option under settings when you type a character into the prompt box it will begin suggesting prompts and you can press tab to autofill the highlighted prompt it will also show you the number of times that suggested tag is used so you can see its popularity the more popular the tag the more likely stable diffusion will recognize it and be able to incorporate it correctly into your generated image while less popular tags will have less references to utilize you also have other features such as the ability to use Auto completion with wild cards extra networks like embeddings and chants which are longer prompt presets you can select which provide a collection of prompts such as starter prompts quality prompts and more if I try to write a prompt you can see it in action and the charts are super useful for filling in detail without doing much leg work the charts are stored in Json files within the extensions tag folder and you can drop in new ones or open the default one and modify it with Visual Studio to create your own chance one button prompt is an extension located within the script drop down of the web UI and this will generate images based on a number of customizable settings you can install this by using the GitHub URL in the extensions install from URL tab before restarting the web UI then you can go to the script option beneath the seeds and in the drop- down list you will see the script called one button prompt the script has multiple options starting with main which is where you select the type of elements you want in your prompt such as the subject artist and image alongside options to overwrite and fter out properties if I run the script you will see how it automatically populates the prompts for us and when the image is generated you can check the generation data to see the prompts which were used then we have the workflow assist tab which will allow to generate up to five random prompts or use your own prompts for testing without having to turn off the extension this can be useful if you have multiple prompts you want to experiment with while still wanting to benefit from the random prompt generation option the advanced tab contains the prompt compounder which determines how many prompts are generated for a single image The Prompt separator where you can determine a symbol and the prompt separator mode determines how the overall prompt is structured the negative prompt tab gives you settings for the randomness of the negative prompt as well as the base negative prompt which you can choose then lastly the one button run and upscale is where you can generate images and upscale them at the same time this section has a number of expensive options so I'll go into detail if I ever covered the extension in its entirety our last extension is unprompted and this is a rather interesting extension that allows the user a programmatic way to interact with the web UI through multiple features such as short codes text to mask body snatcher and reading and writing variables to install you can copy the link into the install from URL option within the extensions Tab and then restart the web UI the extension has two tabs which serve different purposes the first tab on the top ribbon is your un prompted template editor and this is where you can directly modify the templates which you can use to create prompts the second area is the unprompted extension which has a number of accordion for promo resources and Driver run which is used to check for syntax problems but the template section is the interesting option as it allows you to perform interesting functions such as swapping subjects in the image using presets developed by a clip drop and face swaps there's a lot of functionality crammed into the software so I'd have to cover it all in a dedicated video but the last piece of functionality I'd show you is a short code itself if we use the example short code we can see it generates a random person for us the meat of this comes from the template which called a combination of wild cards using the programmatic format to generate images which you can learn more about in their documentation I especially like the range of features included which can assist any workflow from working with regional prompter to making custom templates which can save you plenty of time but if you don't feel like delving into writing custom templates then you can still benefit from the templates provided to generate images effortlessly which can hopefully help your workflow regardless but to wrap things up hopefully these extensions were helpful and if so subscribe like and consider supporting overall on patreon this is Biz genius and I hope you enjoyed
Info
Channel: Bitesized Genius
Views: 4,060
Rating: undefined out of 5
Keywords: stable diffusion extension, stable diffusion extensions, stable diffusion extensions best, Stable diffusion extensions install, stable diffusion extensions μΆ”μ²œ, a1111 stable diffusion extensions, stable diffusion extentions best, stable diffusion extentions install, a1111 stable diffusion extentions, essential stable diffusion extensions, stable diffusion automatic1111 extensions, stable diffusion extension install, stable diffusion gui, Stable diffusion browser
Id: mnp91K7uwfg
Channel Id: undefined
Length: 10min 8sec (608 seconds)
Published: Tue Jan 09 2024
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.