This server is brand new, bear with me as i set things up & feel free to suggest things in #1002293361526460608 !
#š£ļ½announcements
icy forum
icy forum
We've hit 100 members after less than 24 hours! š
cunning rampart
Hello Traveler's!
The link posted below will take you to the beta sign up page, in which we will start sending out invites early next week!.
The webpage and discord server we are making will have the stable diffusion bot running on our a100 cluster pod network. This network is the 10th largest super computer in the world š We have more raw compute power than NASA!
Anyone wanting to use the stable diffusion bot and resulting web apps we have planned for release are encouraged to sign up for beta access here https://stability.ai/beta-signup-form
Hope to see you all in there asap! -Things Subject to change as they happen
<@&1002917600042422362>
cunning rampart
acoustic gust
Beta Signup Form ā Stability.Ai
icy forum
Thanks for 1000 members! ā¤ļø
cunning rampart
1000 members! wow!
surreal pollen
We are making final decisions of which model variant to use for beta, testing some checkpoints
surreal pollen
New checkpoint seems better on short prompts and worse at longer and I donāt know why 
cunning rampart
Report any user sending unsolicited dm's or being toxic to the mods and we will take action. We want to make this a fun and exciting place to hang out and meet new friends, There are active mods keeping an eye on things, please have fun and play nicely.
icy forum
Today is the international "Thank your prompt completion person" day! Be sure to give the buddies that spend too much time copying and pasting strings of text a big hug today.
Once you're done with that, head on over to #1002292387051212860 and grab yourself some juicy new roles.
surreal pollen
Working hard on getting things ready, sneak peek by @supple bone off to the candy land
surreal pollen
New new checkpoint testing seems better at looooong prompts (via HeavensLastAngel)
brittle silo
young jewel
@everyone Starting now, no one will be able to create new invites. Those that are here now have a guaranteed spot in the beta program.
Stability.ai is a community project, and we're thrilled to let you all try this tool before its official launch. As such, through Stability.AI partnership, this is now the official Stable Diffusion Beta server.
Those that get invited via the beta form will be invited into this server to join you all. The bot (@bot)'s primary objective is stress-testing, so we appreciate your understanding for occasional hitches and bugs. Thanks for being a part of the development before the (soon coming) official launch!
icy forum
Hello @everyone! Starting now, the beta signups are being invited into this server.
The emails have gone out to everyone on the beta signup list, so they'll be joining us shortly. Then, sometime this weekend, image creation with Stable Diffusion will start!
We're excited to see you all creating images and sharing your thoughts before long! Don't forget to share with #StableDiffusion and thank you all for your support.
icy forum
Please check both your spam and promotions email folder!
surreal pollen
Dear all,
Thank you for joining the beta programme for Stable Diffusion!
Stable Diffusion is the first of a series of models across image, audio, video and 3D that we are developing as a company and community to unlock creativity globally.
Beta access
The first step to the beta is joining our dedicated Discord here: -
You have 24 hours to use this link before it expires. Please don't share.
When on the server, please be courteous, friendly and follow the rules. If you are not you will be banned and not get to try the great models that come in the future.
After the 24 hours we will activate a bot that you can use to test stable diffusion and we would value your exploration of both the latent space from which it generates images and the outputs themselves.
Feel free to tag your outputs as #StableDiffusion on your socials and use them as you like!
After stress testing the model and receiving your feedback we will move to a website for further tests with some amazing interfaces in the pipeline and a model release, with open source code, data and the model weights themselves.
The Future
We have come together as a group to be 5G for creativity, moving the world from the dial up that we see today.
This will be transformative and are glad to have you along with us.
A billion people creating, not consuming can make a real difference.
However, we donāt have all the answers and are looking for partners, collaborators and those to extend the power of this AI that blurs the boundaries between humans and technology to help us achieve our full potential.
If you would like to use our upcoming API for earlier access to our beta models and make scaling easier, or want to help us meaningfully through partnerships or joint ventures, particularly around content, please do contact partners@stability.ai.
If you want to tell the story of Stability AI and our ecosystem, please do reach out at press@stability.ai. We have a lot of amazing things in the pipeline..
If you are a developer please do join our communities that we will list in the Discord.
For creatives we are trying to figure out how to best coordinate the massive wave of potential this will unleash. Consider yourselves leaders and help us guide this well.
Finally we have an amazing team and are looking to grow rapidly, if you are inspired, join us at careers@stability.ai.
I would like to thank our hard working team and friends who have helped us get here and can't wait to release this and future models to the World.
See you all in the future,
Emad
CEO | stability.ai
brittle silo
You will be able to use the bot <t:1659810600:R>, at <t:1659810600:t> in your local time. https://discord.gg/7aGAZRTe?event=1005411996361052200
icy forum
Last minute opening!
Hello @everyone! For the next 60 minutes, you can invite anyone you like. Bring in your favorite friends, influencers, even family members in! They will all have access to the Beta.
And for all others, the wait is almost over :)
LINK: https://discord.gg/akQHGE4
EDIT: THIS WAS AN INVITE TO THIS SERVER. IF YOU ARE READING THIS, YOU ARE IN THE BETA SERVER.
surreal pollen
30 minutes. Everyone behave and be excellent to each other.
surreal pollen
Dear @everyone ,
Welcome to Stage 1 of the Stable Diffusion beta. There will be three waves. First <@&1004866014133100674>'s will have access shortly, a prompt a minute to start. Then <@&1005191371721822341>'s will get access to a wider range of channels. Finally all our new friends (
) will have access. Click on your own username to see what your tags are. Let's see how this goes!
1. The purpose of this beta is to test variants of the model ahead of release and our infrastructure to scale to make a billion people more creative.
-
We are interested as to what you will find in the latent spaces and how you will push the boundaries of this before it goes out more widely.
-
Post in #1005537874998997045 your thoughts at any time on the model, but please do be considered and thoughtful. We will have special roles for those that help others and take the time to do this properly. Remember this is the first of many amazing models the development team is working on. Audio, Video, 3D and some amazing interfaces...
-
We may adjust the model and parameters to try various iterations ahead of the website and model release.
-
This is not a paid or commercial service, but a test. This model will find its way into all of your favourite image apps and beyond. More features and improvements will be continuously pushed.
6. All of the inputs and outputs are public domain, please bear that in mind. They are not owned by us or you for the purpose of this beta.
-
There will be a guide to running prompts below. Please read it. Bonus points for those that make guides and video explainers, this is a community.
-
You are free to record, make articles on or otherwise discuss whatās in here. Feel free to ping us if you do, we would be interested to see! info@stability.ai
-
Do not be an idiot and just make nice prompts. Not ones that would make kids upset. Or your mother. **Donāt make your mother ashamed. **
10. If you want to make NSFW or offensive things make it on your own GPUs when the model is released.
- Seriously if you do this you hate open source and love closed systems that feed on our attention and productise it. I hope youāre happy.
**12. Zero disrespect will be tolerated, the ban hammer will fall. **If you see anything untoward please report immediately.
Any questions feel free to mail info@stability.ai or ping a mod. Also if you want to help out here, ping us!
Enjoy and donāt forget to tag and attribute #StableDiffusion when you post your masterpieces.
Also donāt forget to ā your favourites and help each other, there is a handy artist guide here by @sureailabs. They get bonus points (points mean prizes): https://gainful-zone-eb1.notion.site/7ddce417501342288b5190aa969549b3
Emad
More features will be added (DM's etc) as we see that everyone behaves. A full terms of service can be found here: https://stability.ai/stablediffusion-terms-of-service
Doing final bot tests. We will have a Center Stage shortly to show how the bot works led by @rotund mica
tepid ibex
Bot is live for early adopters - enjoy š Thanks team.
icy forum
that means - those in the <@&1004866014133100674> role have access, about 1900 of you. We will be expanding after initial stress testing!
Check out the designated #1005512133070557194 channels, at the bottom of your list. Please be patient - the bot's response time seems to be quite slow at the moment. if you encounter unsavory content against the rules, right click > apps > report to staff
Dont forget to share your creations with #stablediffusion, and enjoy!
surreal pollen
All if you need to find your images, go to the top right search bar and type in your username and hit "mentions" that pops up. You get a list of all your posts mentioned including the bot outputs.
surreal pollen
Hi all, it seems that we are actually hitting Discord channel limits at the moment and that is what is causing delays, which is why some channels are faster than others. We are adding more channels and making some adjustments to unblock it. If you feel frustrated go for a gentle stroll or have a nap, this is all very exciting but no rush as this technology is here to stay and will only get better 
Once more - cutting edge AI that can change the World and unconstipate the creativity of humanity?
No problem.
Email mailing lists not hitting spam and Discord Bots not choking?
Weāll sort it š
icy forum
Those not in the early access role are now able to see the channels!
icy forum
<@&1005191371721822341> It's time! Wave two coming in, 5000 new users. Feel free to use the new channels #1005625691146047548 -50! Reminder, if you encounter unsavory content against the rules, right click > apps > report to staff
Have fun, play fair, and use #stablediffusion :)
young jewel
We're currently banning users that are really obviously breaking our rules so please refrain from making prompts that clearly break guidelines you have been warned. We're Watching š
surreal pollen
Easier way to find your images, click the wee mailbox, cheers @regal prism
Yesterday's kick off Stage recorded by @dire sandal if anyone would like to check it out who missed it. Whenever I listen to my voice it sounds so whiny, anyone else find their recorded voice grates? Thanks for listening in everyone https://www.youtube.com/watch?v=EqTLkt-Ycwo
surreal pollen
Hi @everyone
The first test of scaling has gone well. Some people have made their mothers ashamed but by and large has been great.
We will do a daily award for picture of the day, please submit in #1005902846216118342 and vote with emojis for your favourites.
Winners get glory and we will post on our official socials. Maybe moreā¦
We are implementing some additional changes and improvements and will open up wave 3 tomorrow likely around this time of day.
Donāt forget to give your #1005537874998997045 or #1005625710527918110 for those in wave 1 and 2, it is much appreciated.
icy forum
In 24 hours, Stable Diffusion has churned out 1.3 trillion pixels. That's 1,300,000,000,000px ā¤ļø
young jewel
Bot is temporarily down. Please remain patient. This was always a possibility =] I'm sure it will be fine. Just be nice to each other and wait for it to be back up.
surreal pollen
Also adding a few things in
young jewel
bot is about to restart, all in progress dreams will be interrupted, but you can resume having fun dreaming. =]
young jewel
By the way forgot to say the default is now for -n to give individual images, If you still want a grid, use -g
young jewel
Everyone in the server has now been given the beta user role =] have fun. So you can use the bot. =d
surreal pollen
Those who did not get the e-mail to beta testers first has been sent another one, new frens soon. New friends, you will receive access to the beta in wave 4 in a day or so as we absorb the third wave just activated.
surreal pollen
@everyone another wave of beta user sign ups now coming in from those that we expertly hit the spam box of 
Welcome everyone and make them feel welcome in #1006196409046945962. Don't forget to read the #1002292354633453598 and terms of service: https://stability.ai/stablediffusion-terms-of-service
Wave 3 should now have access to the bot, use wisely and distribute amongst channels.
I will announce a winner to the #1005902846216118342 later today, so many great submissions, don't forget to ā or otherwise emote!
As always read the rules, be nice to each other and be excite about some super cool stuff coming 
. Launch announcement for more catchup #š£ļ½announcements message
surreal pollen
There will be a Beta II waitlist going up later today.
surreal pollen
Alright @everyone Beta II final waitlist up, will give it two days properly then maybe no more.
Server holding up well we will look to continue to open up access and please please give feedback #1005537874998997045 #1005625710527918110.
Send your friends that way. Please retweet.
H/t @winter turtle for the awesome healing pics.
Remember to tag #StableDiffusion as you share your creations on your socials!
young jewel
Hello everyone staff have picked some of their favourite for POD posting poll in here. It was really tough picking any of them out. There are to many amazing creations š
deep storm
taps the mic Pardon the interruption, y'all. There may be a small interruption in the near future as we roll out an update. I think y'all will like it. More details when the changes deploy!
The changes have rolled out. The backend service is now massively more concurrent, and will do all the images in a batch request simultaneously. Have fun, y'all!
surreal pollen
Hi @everyone ,
Phew what a few days eh.
You are currently generating 20 images a second, 1.7m a day, or about a terabyte.
Which is crazy, 170 a day per person for the 10k that have access.
We are doing an overhaul of some key systems to make things more robust and so will have to delay wave 4 until tomorrow. Wave 4 is the beta signups which got spammed (sorry) and that closes tomorrow anyway so think that works out, all of you are in the same batch there.
We did one final signup list and will close at the end of today and that will be it. Enough cool people in here, need some QoL improvements ahead of the next phase, the Website launch very
Finally we will be making some (non-public-release) announcements tomorrow, one of the podcasts with the excellent Mat Dryhurst and Holly Herndon dropped today where I discuss more about open source AI, Stability AI and stable diffusion.
Enjoy and let me know what you think: https://interdependence.fm/episodes/open-source-ai-and-stable-diffusion-with-emad-mostaque
See you all tomorrow,
Emad
Ps picture of the day winners soon, will receive an extra special prizeā¦ #1005902846216118342
surreal pollen
@everyone Stable Diffusion is now available for research access: https://stability.ai/blog/stable-diffusion-announcement. Public access to the weights is coming soon as we finish up some things. More later. (link fixed)
Thank you to all I took the pictures from, will thank you individually soon, sorry 
<= this is Emad
Doing a stage at <t:1660154400> to discuss all this stuff and talk about public release of model, model details etc. https://discord.com/events/1002292111942635562/1006926743413723166
icy forum
All dream channels are now available to @everyone ! Enjoy, play fair, and create some seriously awesome stuff.
young jewel
Bot will be back up momentarily please be patient. Glad you are all having a great time
cunning rampart
Bots down, please be patient, working on it š
deep storm
Hot off the transporter pad! The one dimensional being called DreamBotMothership has emerged from its chrysalis and turned into a transcendental six dimensional being! No longer shall we be bound to the shackles of one dimensional bots! Oh, and it can draw pretty ASCII pictures now. And tell you all bout it's internals with tokenization.
More seriously, we have improved the robustness of the bot sixfold, and will hopefully see no more lags or slowdowns. No guarantees that Discord doesn't buckle from the load of more than thirty images a second thoguh!
Rolling out ... now.
IS THAT THE BEST YOU CAN DO? MORE MORE MORE.
Some illustrated examples of our new features:
- ANSI Art: Provide the
-aflag, with-ac textwidthto set the width, and it will turn your image into ASCII art! Try it, and you might see a bit a cool easter egg inside the feature! - Tokenization: Ever wanted to know what Stability Diffusion sees and interprets of your prompts? You can now, with
-tto show you the tokenization of your prompt -- it will show you exactly how your text is broken up and how it will be fed to the AI. It will also warn you when you go over the tokenization limit, which is77, including2tokens for<|startoftext|>and<|endoftext|>.
surreal pollen
@everyone Our amazing friends at Runway ML have been working on dynamic video adjustment, check this out and let us know what you think the future holds 
Text-to-video wen? 
surreal pollen
deep storm
Another day, another update to DreamBotMothership.
- When you give a prompt that is very long and gets truncated, it will tell you where it is truncated, and the returned
!dreamstring will show you exactly what's passed to the Stability Diffusion backend. - It will tell you the actual tokens that are available to you, as
<|startoftext|><|endoftext|>always takes two tokens, so you really have75tokens available. - We do various QoL improvements for you mobile users, such as transforming an em-dash character to dashes.
Also, we made various backend stability and monitoring improvements.
... rolling out ... now!
deep storm
Oh ... sidles back onto the stage and taps microphone I'm sorry to interrupt -- forgot to mention one more thing. When we implemented the concurrency that gave DreamBotMothership its speed boost, images were coming back out of order on the same seed.
Well, that's not true anymore. Given the same -S seed, a -g -n or -g -i will come back in precisely the same order.
So, yes, you can have your cake and eat it too. I've attached a Stability.AI generated cake to share with you all. This cake is not a lie.
surreal pollen
@everyone we kept Phase I Beta invites for those that were spammed open for a few more days but have now closed this Discord ahead of Phase II Beta to new invites.
Congratulations you are the first 15k.
I would like to thank the moderators for their hard work in literally moderating literally millions of images.
If you would like to join mod team or our new community team please email beta@stability.ai
The next phase will be our new website so we can try new interfaces, a new filter so you donāt get accidental oppai and a series of other things.
@The bot will remain open for now and maybe forever, we shall see.
I will do an AMA and stage update this weekend and hopefully introduce some more team members.
Also erm we will award pod creators soon just literally havenāt had time to think.
Thank you all for your support as we make the world a more creative and cool place.
We will make it so you can take all your creations with you wherever you are.
surreal pollen
There is no need for anyone here to sign up to Beta II. You can if you want to for kicks.
surreal pollen
@everyone will do an AMA tomorrow along with some other updates, team getting a bit of rest as it has been, erm, 7 days since we kicked this all off.
Currently testing a whole bunch of things like a filter to reduce unwanted occurrences as well as some stuff not for initial release of Phase II, but maybe not so far after such as..
See you all on the morrow 
surreal pollen
We are going to have a clearer demarcation of official announcements (here) vs community and other things in news. Will also label our socials etc appropriately so you can see what is stability directly š
surreal pollen
Twitter
surreal pollen
Just moving some things around, will send a message here when done, have a walk outside, get some fresh air and natural light after a week of prompting.
surreal pollen
@everyone so we are testing a new classifier to avoid accidental nipples and other things that will be packaged with the weights and some other items so we can release them to the public. The weights will be the same as before and the weights given to hundreds of academics, but this will help guide some of the output in the default software & weight package. We are starting off with high/aggressive settings and will fine tune it down in the coming period until we find the right balance. Please self report anything that slips through or any eye-raising things. Many thanks šš¼
The bottom line is you should see what you want to see with these models no matter what that is. With this, new ethical usage guidelines, bias discussion and other elements released into the public the community will drive forward the various elements so we can all be happy and creative just as we want.
We care deeply about the consequences of releasing something as momentous and potentially societally changing as this and want to do so a responsible and thought out manner but with the belief that people are inherently good not bad and can be trusted. Trust you all, please bear with the blurring and give nice feedback. Thank you, more tomorrow.
surreal pollen
@everyone seeing model instability so will roll back to previous version for a bit, please do not deliberately generate NSFW stuff or you will get banned/are basically being a douche. Think of the poor mods. Cheers.
deep storm
taps microphone Can I have your attention please?
So, y'all having fun with Stability Diffusion? That's good, because we're having fun improving the user experience for y'all. There may be changes that are unpopular with you, but it's in the interests of a streamlined, easier, and smoother experience for all y'all.
-ihas been depreciated. We noticed that a lot of people were being hit in the face by all the generations that people are doing, spanning multiple pages. If you want more than one image, you will need to use-g, in the interest of being kind to our fellow Stability Diffusers here.- However, if you provide
-g, you will receive broken out the individual commands to reproduce the images individually. This makes it much easier to just copy and paste for individual generations. Isn't that neat? As an iPad user, I'm especially appreciative of not having to copy and paste the individual seeds. - In the interests of keeping everyone on an even keel on this free beta service, we've limited the number of steps to
-s 50. Y'all, you're warming the planet for no good reason when you use-s 150on every request with a-n 9! - Consequentially, to bring experiences and expectations in parity with other services, we will be limiting the number of generations to
-n 4per request.
All these changes should deliver to y'all a smoother, easier to use, and faster experience!
deep storm
Content Classification
@everyone taps microphone Oh, I'm sorry, one more thing!
Due to users that cannot keep their fingers and brains to themselves -- this is a PG-13 Discord, for goodness' sake ... we are rolling out the classifiers again. This will reduce the workload on our poor community moderators making sure that we can continue to provide this service publicly and free for y'all while you're beta testing this for us. They are the true heroes around here. You wouldn't want this server to become closed and adults-only by Discord, would you?
However! We hear you loud and clear about the false positives. If you have a blurred generation that you feel like is a false positive and actually would be safe to share with children and families, use the š :seenoevil: emoji to tag the generation in question. Our crack team of analysts will get right onto the problem and incorporate that data point as we refine our first generation classifier. We also have a Report False Positives action that can be taken if you right click or long-press on the message.
Keep in mind that this is not free license to attempt to make the model generate NSFW content. We will still take action on deliberate attempts to circumvent the filters, and appreciate your Report to staff actions in keeping us all open and public. This will also help us refine the classifiers so that less false negatives slip past.
Happy Stable Diffusing, y'all.
deep storm
OK, OK, we know that some of y'all hate the removal of -i. It ain't in the works to put that back. Sorry.
However, you can use a handy utility that @topaz shard coded up to split the images!
Just choose the grid image, put in the size, and boom, it'll split it up for you.
surreal pollen
Twitter
Nearly there will just tweak the parameters here next few days and finish up license then
surreal pollen
surreal pollen
icy forum
You can now split grid images right here through discord to get the individual full-resolution images through DMs!
On desktop, right click > Apps > DM (split) images
surreal pollen
@everyone check it https://twitter.com/Artbreeder/status/1559987252859584512
surreal pollen
Dear @everyone .
As we prepare for the public open beta of DreamStudio and API usage to ramp we need to get everything ready.
As such we are now winding down the discord bot across channels and moving them to read only. You will be able to run your own with API
Thank you for participating in the Beta phase I and giving your feedback.
In the next couple of days we will announce alternate routes to access generations including fingers crossed the model weight releases for V1.
Emad
icy forum
check out #1010300608575709234 for info on the website!
The images will not be deleted, so you may continue to collect your files using the āļø react in the coming days as we move our compute power to the DreamStudio beta website. Images will be able to be migrated over to your website account in due time
rotund mica
Dear @everyone -
Ahead of the AMA beginning in about 40 minutes, we are opening the #1005550141694017536 for you to submit questions ahead of time.
Please vote up questions you'd like answered using ā to help us pick the top questions to address.
If you haven't already signed up for the DreamStudio beta, please do - https://beta.dreamstudio.ai/
@vague wharf is creating Quick Start videos that I highly recommend new users checkout to learn the basics -
We're excited to hear from you and talk with you today.
Bill
surreal pollen
Dear @everyone,
The closed beta website has come under some attacks recently, so we are moving it into safe mode ahead of our public launch on Monday as we tighten up some things and stomp on some bugs.
We've also heard all the great feedback you all have been giving, and will look to implement as much as we can in time for our launch next week - fingers crossed. Keep it coming in #1010312296804917470 , both positive and negative.
Thank you once more; you know we work hard night and day on this.
If you have had any issues please email techsupportds@stability.ai so we can make sure all bases are covered.
The full lite experience should be up and running by launch (touch wood, never launch on a Friday!). The safe mode toggle will be back upon our public launch, so if you need unsafe mode, you may wish to wait until then to generate more images as you will otherwise have lots of blurred pictures. We will share details of some work we have done around this area next week so folk see what they want to see as it were.
I thought it may also be nice to show some of the features that we will add back to DreamStudio over the coming weeks... check out this demo video for a sneak peek.
What do you think? Let us know!
Emad
I swear some folk.. Like go attack somewhere else.
surreal pollen
Stomped on most of the stuff.
Have had to turn off the 200 free credits for new sign ups for now as system was being abused en masse (I really mean en masse) by some folk. How many google accounts do they even have.
Will switch back on on launch or do some sort of invite credit system where you and person you invite get credits yeah.
Figured it out credits back on in a bit šš¼
Fixed new users get 200 credits again. Really donāt go higher on resolution guys unless you know what youāre doing. Those that didnāt get credits contact techsupportds@stability.ai
young jewel
The dreamer community channels are now open! Please stay on topic and make this a constructive part of the discord. My hope for this feature is that you will collaborate and build in these channels and explore many aspects or art together as a community. I hope you all meet wonderful people, collaborate and learn together through what you have made with dream studio.
Just to clarify this is just for sharing what you made not a new bot. see some !dreams already.
icy forum
psst... if you want to see the prompt from one of the images you see, click on it > open original, and look at its filename...
surreal pollen
Twitter
icy forum
and with this, the new corner category has opened! from #1011228642472378369 to #1011228477954998273, there should be something for everyone.
This server is no longer private, so feel free to share the link: https://discord.gg/stablediffusion
fluid jungle
Stable day to you all! A lot of questions are running rampant as we have finally officially released DreamStudio! That being said, there are also a ton of questions and help you may have and need! The staff have implemented a ticketing system here: #1010934719455707218 where you may forward your questions to our staff and mods to help you shortly! Thank you for your patience and support on our big day! Let's dream!
surreal pollen
@everyone would also like to announce DreamStudio goes public, share around. I would also like to express my appreciation to each and every one of you, thank you and please be patient in the coming days as things go a bit crazy, everyone is doing their best. Help each other out. https://twitter.com/EMostaque/status/1561777997748862976
lyric bobcat
LAUNCH AMA with @surreal pollen 
@everyone
https://discord.com/events/1002292111942635562/1011369206454960198
surreal pollen
Twitter
surreal pollen
Twitter
deep storm
@everyone For those of you who are wanting to work with Stability.AI's image generation service via API, we have a public repository (https://github.com/Stability-AI/stability-sdk) with the following:
- A reference Python client that can be installed via
pipusingpip3 install stability-sdk - A
ipynbnotebook that shows how to use it fromipythonor Collab (https://github.com/Stability-AI/stability-sdk/blob/main/nbs/demo_colab.ipynb) - A
.protofile that can be used to generate gRPC (https://grpc.io) stubs for your favorite language! We have internally used this to generate stubs for Python, Golang, and TypeScript/JavaScript. (https://github.com/Stability-AI/stability-sdk/blob/main/src/proto/generation.proto)
Right now, this requires a bit of "assemble before play", but we are working on making the stub generation more automated across multiple languages.
More will be coming soon, including an open sourced high-performance Discord bot written in Golang.
deep storm
@everyone I thought y'all would like to see a WIP snapshot of the next version of the Stability.AI Discord bot v2 that we are intending to open source. It's a much nicer interface than the old bot with slash commands, and buttons to manipulate the image. The only manipulation implemented right now is Redo, but there will be other ones such as changing the step count, the CFG values, or the size. You can even cut and paste slash command!.
This open-sourced bot will be able to take your Stability.AI API key that you can get from your account at the Dream Studio website, and it will use your credits there to generate images in the comfort of your own Discord server or community.
surreal pollen
Twitter
deep storm
It's time to perform a pip install --upgrade stability-sdk!
We've made an update to our stability-sdk to v0.1.1. The most important change is that we removed the dependency on python-magic. This means that those of you that are on VPS or notebook hosting providers that do not allow system packages to be installed will no longer be blocked on installing the magic binary package.
There's some other minor changes in the code, such as variable cleanup, but these should not affect you for the most part.
NOTE: In the future, these sorts of code release and SDK related (š¤) announcements will be pinging the <@&1004114999867879506> tag. If you want to keep abreast of these updates, go over to #1002292387051212860 to opt into the <@&1004114999867879506> .
surreal pollen
@everyone first announcement as we move to create an intelligent internet where our AI brings all content, public and private, to life.
Eros is an amazing partner (Bollywood models incoming..) and we will make it so India gets this technology years before they otherwise would to help them accelerate even more.
Much more to come.
Twitter
The announcement also teases some of the other unannounced models and technology we have coming, see if you can spot it ^_^
surreal pollen
Peering at v1.5 (by @formal swallow)
icy forum
We welcome people from all over the world ā¤ļø š š š
deep storm
@everyone Stable day and stable night to you all! We present to you, our community, the 1.5 model update for all of you to test for a limited time! Please be sure to leave your input of this update in #1010312296804917470 channel and get back to happily /dreaming!
Along with the 1.5 model update is a brand new bot rewritten from scratch; we affectionately call it GoBot because it is written in golang. It is much faster, and more more user-friendly using slash commands! To access it, use /dream in the #1013576650434297937, #1013576694734520380, #1013577322747674695 ... #1013578413015052399 ... #1014333048118050846 and so on channels!
One of the nicer features is the aspect ratio feature. Rather than trying to go for a specific size, I would suggest that you use aspect to get the ratio you want.
- Given no
height, orwidth, it is done as a calculation of the square of512. - Given either
heightorwidth, the other dimension is calculated as a ratio off the provided one.
Any dimensions you have must be equal to or less than 294912 pixels.
One piece of good news -- no more slow-mode! The bad news is that you can no longer simply surf the channels, as the bot will put you in the penalty box if you attempt to do concurrent image generation requests. Repeatedly clicking will cause your penalty box time to increase exponentially as an anti-spam feature.
Happy /dream'ing!
fluid jungle
@everyone Hello hello! We see you all have been /dreaming away with some very interesting and spectacular outcomes! As mentioned in our previous announcement, this would be a limited time action! This being said, we have gathered the data that we needed from and about the GoBot -- and will be winding down the #gobot channels. 
The GoBot has served up nearly half a million stable-diffusion-1.5 images in less than 24 hours, while never breaking 10% utilization of a single CPU core! We've collected feedback and will improve the GoBot code, before making an Open Source release of the fastest and most performant image generation Discord bot around.
fluid jungle
@everyone 
There will be an exciting announcement and AMA with Emad and special guest Anjney Midha happening in the Developer Lounge Server <t:1662140700:R>! Come join us soon with all your burning questions! discord.gg/BJ4JRfTTFt
lyric bobcat
AMA SATURDAY @everyone
Stable day to you all! Load up your questions for #1005550141694017536 which will be open just before our AMA with <@&1008434998698840124> on Center Stage. Check event details for more info. See you dreamers 
https://discord.com/events/1002292111942635562/1015355035313504397
rotund mica
@everyone
Stable Diffusion v1.5 is now live inside of the DreamStudio website - https://beta.dreamstudio.ai/
Yes! The very same v1.5 that you helped us test here inside of the Discord Server last week in the GoBot.
And yes, there is an option to switch back to v1.4 with a dropdown selection located on the righthand config panel.
We'll be rolling out more updates this week to the DreamStudio beta website.
Keep checking back.
Thanks!
Bill
surreal pollen
Version 1.5 is not currently available for download but will be in due course after we put some more procedures in place.
This is likely to take a week or two.
We are balancing how to communicate a roadmap better while not pressuring the developers.
1.5 is the same number or parameters and size as 1.4 but has further training as we look for an optimal balance, it should give better results.
More details will come with the upcoming DreamStudio feature update.
deep storm
<@&1004114999867879506>
A reminder that all of Stability's models are available via API, and that the code can be found at: https://github.com/Stability-AI/stability-sdk
For those of you who use the stability-sdk and the API, you can switch to stable-diffusion-v1-5 in the API.
For Python, either instantiate StabilityInference with the engine="stable-diffusion-v1-5" argument, or set it on an already running instance: stability_api.engine="stable-diffusion-v1-5"
For JavaScript/TypeScript:
request.setEngineId("stable-diffusion-v1-5");
icy forum
Hey everyone,
Emad will be hosting an Ask Me Anything event on Reddit this Friday!
If you haven't been able to get your questions to him during live stage AMA's, this is your chance to ask to your heart's content over a text format.
Join us! We'll be posting the link to the post here <t:1662728400:R>
rotund mica
**Image to Image / Initial Image now live on DreamStudio beta
**
@everyone
We have a new image creation feature in DreamStudio beta, some UI updates, and weāve improved the sampling / reduced the amount of noise overall with images in DreamStudio and via our API.
The new image creation feature is known as āImage to Imageā (or āInitial Imageā) and it allows users to upload a picture as the input source for Stable Diffusion to use in generating new images.
In combination with text, the strength or āimage influenceā of the uploaded source picture can be adjusted with a slider to adjust the intensity of the uploaded imageās impact on the resulting new image.
It sounds complicated, but once you see it in action, I think youāll get it ā and once you try it, youāll likely be obsessed with it like I am. Speaking of seeing it in actionā¦ Yes!! Weāve got two short video tutorials (by Zippy & KaliYuga) to get everyone up to speed and so you can see it in use before giving it a try. Links below -
Getting Started With DreamStudio Website Beta, Part Four: Image Guidance
https://www.youtube.com/watch?v=PnYM6Wym_Ck
DreamStudio Beta - New Image-to-Image Feature!
https://youtu.be/M3jCa6qTHpQ
We hope you find these tools valuable and enriching and we thank you for being part of the Stable Diffusion and DreamStudio communities.
Bill
deep storm
Howdy, y'all <@&1004114999867879506> s!
We heard you, especially y'all TypeScript/JavaScript developers. We'll get to that in a moment.
First off -- our stability-sdk [1] has been updated with the latest API methods to do init images (aka img2img) and masking!
But almost more importantly, we've opened up our internal api-interfaces repository [2]. This is our source of truth for interaction with the Stability.AI API. You now have the same level of access and information on the gRPC-based API as our internal developers do.
What does all this mean? You can use the API to generate images from prompts, including init images with masks! So if you have something programmatic, and you wanted to use the thousands of GPUs available for your project from your notebook or application, you can!
It includes:
- The original Protobuf
.protofiles andCMakeFilethat we use to drive regeneration of our stubs. - Pregenerated Python client/server stub bindings (but it is recommended that you use the
stability-sdkreference client due to Python's lack of static typing) - Pregenerated
golangclient/server bindings - ... and last but not least, pregenerated TypeScript/JavaScript client/server bindings!
We have gRPC interfaces defined for:
generation-- the main API y'all will be using to generate artifactsengines-- to enumerate the available engines/models from the APIdashboard-- if you wanted to find out how much credits you had, and interrogate the state of your accounts.
In other news, work on completing gobot on the way to open sourcing it is progressing well. It has init image support, and supports interacting with the init images. Next up is adding support for engines and dashboards so that you can check your balance, and enumerate the models available. I've attached a sneak peek video of gobot in operation.
[1] https://github.com/Stability-AI/stability-sdk
[2] https://github.com/Stability-AI/api-interfaces
surreal pollen
@everyone want to use Stable Diffusion in Photoshop with your API key from DreamStudio? Well you can now! Great work by Christian and more to come! https://twitter.com/cantrell/status/1567898557763538947?s=20&t=0xS-PM0i8eyPmLXuYzP9yQ
lyric bobcat
THANK YOU FOR 50K DREAMERS! 
šµ
šØ @vague wharf
icy forum
...Do you see that green badge?
This server has been verified as the official Stable Diffusion by Stability AI Discord server! 
Thanks to everyone for joining us on this journey. Dreaming of the future 
surreal pollen
@everyone Happy to announce new OpenCLIP models after 1.2 million A100 hours of effort.
These models are embedded in stable diffusion, DALLE2 etc so you will see another leap in quality and are used to guide MidJourney and others which will also improve.
Previously only the smaller CLIP models were made available by OpenAI who kicked off this revolution by doing so, happy to make the larger, better, models available to all.
https://twitter.com/emostaque/status/1570501470751174656?s=46&t=jTh68A_YCxdzuaOxZyJB5g
surreal pollen
@everyone 1,000,001 https://twitter.com/EMostaque/status/1571572181976465417
Twitter
rotund mica
In-Painting / Out-Painting now live on DreamStudio beta (via Chrome)
@everyone
We have another new image creation feature available now inside https://beta.dreamstudio.ai/
In-Painting / Out-Painting tools are live for DreamStudio beta users via Google Chrome. We are working on adding support for more browsers in the coming weeks. For now, we hope you give it a try on Chrome. Please note: in-painting and out-painting are currently only available on desktop/laptop computers via the web app. Mobile availability is TBA.
Weāve also made some important UI/UX updates, and weāve streamlined the Support Request system for a faster and more organized way to submit feedback on the beta.
In-Painting/Out-painting allows users to upload a picture as the input source for Stable Diffusion, use a brush to erase sections of the source image, and then send that image back to the model as the source for the new image.
In combination with text, the strength or image influence of the uploaded source picture and the masking settings can all be adjusted with sliders to change the intensity of the uploaded imageās influence on the resulting new image.
Same as anything AI when explained solely with words, it may sound complicated, but once you see it in action, youāll get it ā and once you try it, youāll likely be obsessed with it like all of us at Stability are. As always, weāve got a short Quick Start Guide by @vague wharf to get everyone up to speed and so you can see it in use before giving it a try.
Getting Started With DreamStudio Website Beta, Part Five: https://www.youtube.com/watch?v=GRXrzC76w6o
We hope you find this new toolset valuable, and we welcome your feedback via the Support page inside of the beta website, here in #1010312296804917470 and be sure to check out our updated FAQ
https://beta.dreamstudio.ai/faq
Thanks again for being part of the Stable Diffusion and DreamStudio communities.
Bill
fluid jungle
Dear dreamers, drumroll please š„ *:..ļ½”oā
We present to you our first winner of the most highly rated image in our starboard server-wide! Brought to us by @torpid sonnet , they created a beautiful selection of colour themed Ariana Grandes! Beautiful, right? 
Big shouts to them and Congratulations! 
We also have some exciting events starting next week with some cool prizes! If youād like to be pinged for weekly challenges and events moving forward , please opt in to our challenger role with āļø over at #1002292387051212860 Also be sure to head on over to #šļ½winner-gallery to have a look!
Thank you so much for all your patience and we look forward to seeing your creativity in motion soon! āļø @everyone
heady fractal
Hey folks! Got a minute to take a fun, super-quick survey (just 2 questions)? The Stable Diffusion dev team is working on some cool new features, and we could use your help to figure out how best to describe them, so they make the most sense to you & others like you. Cheers and thanks in advance; really appreciate everyone's feedback! https://forms.gle/GupaQTuQW1Uct7cu8
fluid jungle
<@&1018216027244466246> Stablest of Diffusers, we heard you were looking for some challenges?
Weāre happy to announce our Community Events Center will be launching over at #1023999442338201721 ! You can submit your art based on curated topics, get involved in some unique prompting challenges, and compete for weekly prizes! All the info youāll need will be over there waiting for you~
Hope to see you there ! āļø
fluid jungle
Dear Dreamers,
After having the pleasure of seeing all of your ethereal wonderland submissions, we are thrilled to announce our very first POW winners for Week #1!
MASSIVE congratulations to
@proven timber for scooping up the #1 winning spot with this ethereal masterpiece "World of Glass Bubbles"!! Youāre now gifted with your art becoming our new Server banner, wielding the Stable Society Role, DS Credits, Discord Nitro, and shoutout on Socials!
In 2nd place, we have @viral owl with a rosy delight wonderland 
And in 3rd place, we have āA slice of paradiseā by @covert crypt š°
The energy and group effort you all have made has been so wholesome and heartwarming. Thanks so much for kicking these events off with a š„ ! Be sure to take a peek at #1023999442338201721 to see all the beautiful contributions to this weekās theme, and check out our #šļ½winner-gallery to see all these pretty pieces from here on out.
If you missed out this week, be sure to opt in at #1002292387051212860 for weekly art challenges!
Stay tuned! <@&1018216027244466246>
fluid jungle
šÆļø Dark Diffusions - The Monster In MešÆļø
<@&1018216027244466246>
After having such a wonderful response to our last POWš„Event, we present to you for the month of October: 
š» Dark Diffusions! š»
Letās dive into our first theme brought to us by our dear fellow maven @flint osprey š¦ 
The Theme for Week #2 is āThe Monster in Me" so let's unleash our imagination! Before you start creating and experimenting with your prompts, let's take some time to think about how youād like to portray this topic.
What kind of monster do you want to create?
Where will this monster be?
Are there any other elements that are important to have in your scene?
NOTE: (Not all monsters are seemingly scary, right? Some can be quite abstract, some represent concepts much bigger than us like our hopes and fears, and some may even appear harmless all together.)
How would you like to express your Monster? How does your Monster express you?
If you want to start by creating something similar to our monster in our flier, our base prompt was:
A photo of a halloween anthropomorphic monster walking on a forest, concept art by justin gerard, an album cover by jean-baptiste monk, a detailed matte painting by christophe vacher, a detailed painting by fyodor rokotov, d & d, a character portrait by todd lockwood, behance contest winner, fantasy art, concept art, storybook illustration, lovecraftian, zbrush central contest winner, furry art, reimagined by industrial light and magic, gothic artš¦
Tried it? Wonderful! Good to experiment, right? But Iām sure that you can do much more and deliver even more alluring outcomes!
Head on over to #1023999442338201721 to start your journey and see all further guidelines and head to #1002292387051212860 to grab the "challenger" role to opt into future events announcements! 
fluid jungle
Hi there! Friendly reminder that voting for all Monster In Me submissions over at #1023999442338201721 will be closing <t:1665086400:R> Be sure to vote for your favorites! 
fluid jungle
šÆļø <@&1018216027244466246> šÆļø
**This week's POW #2 WINNER **for the theme - The Monster in Me
We asked you to deliver us your spookiest and most thoughtful submissions you could give, and man, did you guys deliver! From jarring and unsettling to clever, it was a grab-bag of tricks and treats!
There was one user in particular who prevailed in submitting an absolutely beautifully bone-chilling entryš¦“
DRUMROLL
Let's give it up for @ripe light and their emotion packed piece for our theme: The Monster In Me. Nerf's entry will now be our new spooky banner for the week and partake in some epic treats!
Go check out their submission over at #šļ½winner-gallery and show it some much deserved love! 
If you'd like to follow along with next week's POW, skip on over to #1002292387051212860 to pick up the "challenger" role!
surreal pollen
@everyone it has been a while. It is 11:28 pm in Geneva now and am with Red Cross, day after day..
Due to the publicity of the recent incidents, we wanted to remind everyone of some rules of the Discord server.
Please understand that while our own goals for Stable Diffusion promote open source & community participation ideals, we do not support the use or spread of code or models that were illegitimately obtained or distributed.
We do not support any project that promotes this and would urge caution.
You may discuss this situation in a respectful manner on this Discord but please keep things civil.
I will send more thoughts soon, we have been deeply engaging in various elements around this and other topics.
We will also have a AMA on Monday to discuss and get input on various topics.
Thank you,
Emad
Also working hard to get new models out, think processes fingers crossed nearly done. Working hard on new datasets and have been releasing models and code in other modalities in the meantime. Will discuss more Monday.
Will post an agenda this time. Pardon the typos am kinda exhausted.
surreal pollen
One final note.
We worked with the team at NovelAI to build some of the scaling stuff around SD and more, their input has been invaluable.
Donāt take anything out on mods etc if youāre pissed, buck stops with me and I made some decisions after careful consideration.
Look forward to speaking with you all Monday, been too long.
7 weeks -2 days inā¦
fluid jungle
<@&1018216027244466246>
POW Week#3- "House of Solitude" Winner Announcement
The wind has finally come to an abrupt haltā the leaves have settled to a gentle crunchāš
Just enough time for us to announce our POW winner this week for Dark Diffusionsš¦
š„³ CONFETTIš
BIG round of applause to @arctic leaf for their intricate submission of a temporary solace! What a great way to show that even disorganized and non traditional methods of comfort are still effective!
We also had our first ever TIE in POWā which resulted in a good olā fashioned staff-pick for the winner! Weāll be gifting both of these submissions with the Stable Society role this week, so big shoutout to @willow fossil for their unique piece as well!
Thank you all for your positive energy and life youāve brought into these events!
Until Next week~
Take care!
Go check out their submission over at #šļ½winner-gallery and show it some much deserved love!
If you'd like to follow along with next week's POW, skip on over to #1002292387051212860 to pick up the "challenger" role!
last jackal
Hi @everyone
We're excited to release a significant improvement to DreamStudio!
DreamStudio will now use CLIP guidance to enhance the quality and coherency of your images, improve inpainting/outpainting, and give much better results for complex prompts.
This is the product of weeks-long tuning of settings across a wide variety of image types. We have also put in place several other image enhancements, and we have adjusted the minimum steps to 35, to assure consistent results across all image settings.
We hope you'll agree that the new images are amazing! (Some awesome samples down below)
If you prefer to use DreamStudio without CLIP guidance, just turn it off with the toggle switch. There's no additional cost to use CLIP guidance.
This upgrade is part of our ongoing beta test, and we welcome your comments.
Please let us know what you think in #1010312296804917470, and thanks for Dreaming with us!
ripe torrent
Happy Saturday all! Since we've heard some people depended on it, we've enabled the ability to go below 35 steps when CLIP guidance is disabled. Simply toggle CLIP guidance off and the slider will update with the extra range. Thanks for your feedback on this!
night dove
I'm Daniel Jeffries, the CIO of Stability AI. We've been heads down building out the company so we can release our next model that will leave the current Stable Diffusion in the dust in terms of power and fidelity. But because we've been quiet that leaves a bit of a vacuum and that's where rumors start swirling, so I wrote this short article to tell you where we stand and why we are taking a slightly slower approach to releasing models. The TLDR is that if we don't deal with very reasonable feedback from society and our own ML researcher communities and regulators then there is a chance open source AI simply won't exist and nobody will be able to release powerful models. That's not a world we want to live in.
https://danieljeffries.substack.com/p/why-the-future-of-open-source-ai
fluid jungle
<@&1018216027244466246>
**POW Week #4 - āA Ride Into The Unknownā Winner Announcement **
Hello hello! Look how far our journeys have taken us this week~
We saw cars made of fries, portals to the other side, and trains with blinding lights-
You all have been nothing short of creative, that's for sure!
Letās buckle up for our NEW POW Winner this week for Dark Diffusions, shall we? š
WHOOSH
Round of applause to @dense kestrel š š„³ for their lovely submission of an otherworldly train that exudes mystery and the excitement of unknown destinations.
Go take a gander yourself over at #šļ½winner-gallery and give it some emote love!
If youād like to follow along with us for next weekās POW, go on over to #1002292387051212860 and select the āchallengerā role!
fluid jungle
Stability.ai Update! @everyone
We had our official launch event on October 17th and a lot has happened! 
Watch the full recording on our brand new Youtube channel!
(like and subscribe or 
) https://youtube.com/watch?v=S3qlqY_sOPw
šSome of the exciting projects in the works and announcements made: š
šļø Our CEO Emad joined the NewYorkTimes podcast, Hard Fork, to discuss Stability AI's approach to open-source AI and share some of the recent announcements.
DreamStudio, our AI creative media platform at https://beta.dreamstudio.ai/ reached 1.5 million users in 7 weeks! š
We also showed a preview of DreamStudio Pro, our upcoming offering with animation and a powerful integrated suite of creative tools - See its mind blowing capabilities at 23:48.
We announced a partnership with Leia Inc., the leading 3D experience platform leveraging nanotechnology and AI. *I kid you not, it's insane! * 3D visuals on a tablet without special eyewear! This will bring Stable Diffusion content in a 3D immersive experience.
š¤ We announced a partnership with Eros Investments to bring Stable Diffusion to India and to collaborate on Immerso, a new interactive "virtual universe" powered by ancient mythology, an innovative intelligent eco-system, and Stability.ai!
š¤ Stability.AI supports various open-source AI research organizations also with exciting announcements:
- CarperAI recently announced plans to create the first open-source "instruction-tuned" language model to improve safety and usability. You can discover more here : https://twitter.com/carperai/status/1582891780931874819
- OpenBioML announced LibreFold, a project focusing on improving AlphaFold-like methods.
They tell you more in this post : https://twitter.com/openbioml/status/1582401316961165314
And much much more cool stuff in the works at: - EleutherAI : large language models & AI alignment research https://www.eleuther.ai/
- LAION : making large-scale multimodal AI openly accessible https://laion.ai/
- Harmonai : open-source generative audio tools https://www.harmonai.org/
Stay tuned, here and through our socials--- More announcements coming !
Together, we'll help create a brighter future by building open AI tools to unlock our potential and have lots of fun along the way.
š Thank you all for your support, creativity, and good vibes - great communities are what fuel this rocket!š 
Check out some of the highlights here! https://twitter.com/StabilityAI/status/1585047427136229376?s=20&t=8yP7WsPrj9i7TpCudclxVQ
fluid jungle
š£ POW Week #5- āMovie Night Frightā Winner Announcement <@&1018216027244466246>
The curtains on Dark Diffusions have come to a close! We made monsters, creepy houses, crazy cars, and some pretty phenomenal movie posters this month!
For our next month of POW, we will be diving back into a grab-bag of different sorts of open-ended prompting challenges for you to take part in! Thanks so much to everyone who participated and made this event such a cool little nook on this server. 
**Without further ado, **
Letās give it up for our full circle winner @proven timber !!!! 
š for coming in with an entire RECAP for this event! Drite was our very first POW Winner, and has sealed up our spooky season with this incredible submission. Congratulations!
Iād like to also give a warm shout-out to all of these lovely users for their great submissions and their equally great vibes during the entirety of these events: @real wolf @vague bobcat @thick minnow @light shuttle @winter viper @blazing knot @wooden lava @hybrid urchin @misty prawn ! 
Go take a look over at #šļ½winner-gallery and give our winning submission some emote love!
If youād like to follow along with us for next weekās POW, go on over to #1002292387051212860 and select the āchallengerā role!
fluid jungle
**POW Week #6- āTime Capsule Memoriesā Winner Announcement ** <@&1018216027244466246>
Memories, memories, memories
We were able to sift through the time capsule you provided to us and uncovered our nostalgia winner for this week!
ROUND OF APPLAUSE
Congratulations to our WINNER @boreal dust !!! for transporting us back in time to their distant and fond memories of being a kid again!!!š 
And congrats to our TIED runner ups! @proven timber and @scarlet flicker 
Go take a look over at #šļ½winner-gallery and give our winning submission some emote love!
If youād like to follow along with us for next weekās POW, go on over to š§ #1002292387051212860 and select the āchallengerā role!
BONUS: little gift (from us, to you!) for all of our participants: https://tinyurl.com/34brrs3u Typically we remove images after our challenges are complete, but this one felt very special and close. Thank you for sharing all of your treasured memories with us. Let's continue to create more of them together.
fluid jungle
POW Week #7- āAbstralitiesā Winner Announcement <@&1018216027244466246>
This has to be my very favorite POW thus far! The submissions you all showed us were some of the very best! Broken down realities, surrealism, escapism-- they all mingled together to create something truly amazing.
Our winner this week is our very own @proven timber , with a whopping 25 points!! š„³ š šŖ© Congratulations!
Closely behind, our Mod's @forest field and @winter turtle tied for 2nd place! W's all around!
Go take a look over at #šļ½winner-gallery and give our winning submission some emote love!
If youād like to follow along with us for next weekās POW, go on over to #1002292387051212860 and select the āchallengerā role!
We'll see you next event for something very......š š š š
fluid jungle
š **POW Week #8 Challenge- "The Never-Ending Loop" Winner Announcement & Updates!**š <@&1018216027244466246>
You all sent us into a...spiral(hold your applause, please) this week with all of your incredible submissions. It's always so interesting to see how the community interprets open ended sorts of writing in their own ways and unique styles! This week we add a new member to Stable Society !
Give it up for @tacit delta for their paint-swirling infinity Winning Submission! Congratulations for an incredible entry! 
š
Our runner's up ...all locked into a 3-way tie just a point behind! Shoutout to @earnest whale @cold sparrow and @proven timber ! 
Go take a look over at #šļ½winner-gallery and give our winning submission some emote love!
If youād like to follow along with us for next weekās POW, go on over to #info and roles and select the challenger role!
UPDATES
This weekend, we will be rolling out Dailies to our Community Events Center as a treat for you all! The channel will consist of our bot giving a random topic per day that you all may submit as many images as you like and riff off of one another idea's together. This event will not entail prizes or specific rules to adhere by, but will serve as a in-between casual event for users to participate in daily, while in wait for the next POW! Keep your eyes peeled whilst more info is underway š !
We are actively collecting your feedback from POW. (Many thanks to those who have filled out the survey thus far! )As we navigate through the best ways to encourage majority's voice through these challenges, we wish to have these challenges continue to be a cozy nook for all of our members and nurture an encouraging environment. We value your feedback and thank you for your patience as we gather your data on the best way to move forward with certain aspects of our contest. Please go here to leave your thoughts: (https://forms.gle/ZhaDuKZnWuTRhCnA7) 
fluid jungle
<@&1018216027244466246>
Todayās the day we roll out #šļ½dailies to the Community Events Center!
Our bot, at <t:1668963600:t> everyday will throw out a random topic for you to generate an image for. Dailies will not have any rankings, prizes, or outlines similar to POW, but will serve as a fun mini-game of sorts to test the limits of your prompting, share your prompt building, and utilize your imagination!
To go see todayās topic, head on over to the #šļ½dailies channel and check the pins, or run the command /topic !
Have fun! 
fluid jungle
PSA: Please be careful when downloading model checkpoints. Treat them as you would any other random executable that you may download.
Due to the nature of open source software, security risks can occasionally arise. Checkpoints can be infected with executable code, and while some hosting platforms such as Hugging Face are implementing scanners to help mitigate this potential security risk, always be sure to do your due diligence with regard to where you are getting your models from. See here for more information on Pickle attacks and risk mitigation strategies from HF: https://huggingface.co/docs/hub/security-pickle
Official Stability AI model checkpoints will always be safe, and accessible under Stability AI's HuggingFace page at: https://huggingface.co/stabilityai
surreal pollen
@everyone
Dear Diffusers, Drum roll please
We're incredible excited to announce the open-source release of Stable Diffusion 2!
Stable Diffusion V1 led by CompVis changed the nature of open source AI & spawned hundreds of other innovations all over the world. We hope V2 also provides many new possibilities with these incredible big improvements and features listed below
So Whatās New :
-
Stable Diffusion 2.0: An all-new text-to-image model trained with a brand new text encoder OpenCLIP, greatly improving the quality of generated images relative to earlier V1 releases
-
Updated Inpainting Diffusion: A new text-guided inpainting model fine-tuned on Stable Diffusion 2.0
-
Upscaler Diffusion: Enhance image resolution by 4x while preserving fine details
-
depth2img: A variant image-to-image model focused on the overall structure and shape of input images, allowing you to radically change up the contents of your images without altering their composition
For more in depth details on the latest Stable Diffusion 2 Release, please see our full blog post
here: [https://stability.ai/blog/stable-diffusion-v2-release]
Note: If you own or maintain an application that uses Stable Diffusion, you will need to make a few small changes to your appās configuration in order to support version 2.0. For more details about accessing the model, please check out the release notes on our GitHub: https://github.com/Stability-AI/StableDiffusion
Shall be hosting a Q&A tomorrow so we can chat about fun stuff. See you all there š¦ https://discord.gg/TeTtZGTq?event=1045032204557897768
fluid jungle
POW Week #9 Challenge- āImaginary Friendā Winner <@&1018216027244466246>
What a lovely wrap up to the week! 2.0 and our Weekly POW winner! 
Everyone make some noise for @winter turtle and his magical capture of this weeks theme!š 
First Mod winner! Congratulations!
Go take a look over at #šļ½winner-gallery and give this submission some well deserved emote love! 
SIDE-NOTE: Thank you all to those who have submitted their POW Feedback! Please be patient with us as we compile all of your feedback and thoughts on the ins and outs of POW. Any and all updates/changes that occur with POW will always be communicated to you in the form of an updates post in the near future! Thank you so much
lyric bobcat
š
fluid jungle
@everyone
Hello hello and Happy 100k Decemberrrrr! 
This month, as our gift and thanks to you, weāve got quite the stack of News, Updates, Nitro, Credit giveaways and surprises in store to round the year off. We hope you can join us! ā
Here's our lineup menu:
Return of Dreambot
No more āWHERE'S THE BOT?ā Welcome Dreambot back with loving arms!
Get ready to /dream again real !
__ POW Week #10: ANIME__
December 6th-12th, 2022
As a nod to our largest dreamer community, we will be having a week-long POW solely for the best anime characters you can generate! (yes, bring the waifus.)
Stable Stage Presents: TNR with Black Label Art Cult & Amli
December 8th, 2022
POW Week #11: Best of SD
December 13th-23rd, 2022
Another time-extended POW! What better way for a Finaleā? Bring your very best generations youāve created this year and show us how far youāve come battle it out for the top spot!
__Stable Stage Presents: Q&A with Harmonai __
December 15th, 2022
Ask all your burning questions centered around AI sound and music generation
TWITCH LAUNCH! 
š āļø
https://www.twitch.tv/stabilityai
deep storm
š¤ @everyone š¤
Dreambot Returns!
From the ashes, Dreambot has risen and returned, with the role taken on by gobot in advance of its open source debut!
It is free to use and access from the Stable Diffusion Discord, and uses the higher quality 768x768 resolution Stable Diffusion v2.0 model. We are excited to bring this demonstration of the model's capabilities here.
To get started, head over to one of the many #dreambot channels, and type /dream. The various options will pop up if you hit tab on a PC, and will be available as buttons on mobile.
Dreambot Features
The gobot code is the most advanced, featured, and fastest AI image generation Discord bot out there. It is able to do:
Aspect Ratios
It is strongly suggested that you use the aspect feature to select an aspect ratio than to use manual height or width setting. gobot will manage the calculations for you, especially if you set one dimension.
Variations
Not quite happy with an image, but don't want to completely redo it with a ReDream? Use the variations feature!
If you generate an image, below the image you will see various buttons that represent the seed number of corresponding images. Clicking on the seed number will cause variations of that image to be created.
Image to Image
If you select the init_image option, the bot will also do image to image generations for you. You can adjust the image_strength from 0.0 to 0.95 to change how strong the image is vs the prompt you give it.
Outpainting
If you have an image that's chopped off or not framed correctly -- or want to change the aspect ratio of the image? No probem, use the AI Outpainting feature, available in the AI Outpainting dropdown.
It will show various outpainting resolutions, and aspect ratios that you can select. Once selected, it will perform an 'outpaint', and expand the edges of your image.
If you look at the available command, you will also find outpaint_* suite of options, includng outpaint_direction if you want to do it on an image that you upload for image to image.
Attached to this post are some examples! You can make some very neat zooming out animations, or extend someone's body.
Guider Prompt
By default, gobot uses CLIP guidance to enhance the quality of the image. By using guider_prompt you can steer the image in interesting directions. You can also disable the guider by providing False.
Interrogate
If you forgot what prompt and settings went into an image, and you saved it correctly -- you will be able to use /interrogate to look up the metadata about the image!
Happy Dreaming!
forest field
POW Week #10 Challenge- āStable Kitchenā Winner <@&1018216027244466246>
Thank you all for making us hungry with your delicious submissions and participating in our Weekly POW!
Stable Kitchen was brought to you by the awesome @flint osprey and his prompt: Delicious big burger and french fries.
Feast your eyes on this week's submission from our POW chef š š @chilly flicker š š with their incredible generation that captures the essence of our theme and cravings. š„Ŗ
Head on over to #šļ½winner-gallery to check out their submission! 
Make sure to grab your challenger role if you haven't already over at #1002292387051212860 to be pinged for our future events!
We'll see you next week for POW: Anime, looking forward to seeing your waifus #1010577750077210726 š
deep storm
š¤ Stable Diffusion v2.1 Preview on Dreambot š¤ @everyone
In advance of the Stable Diffusion v2.1 open release this week, we are bringing you all a sneak preview using the Dreambot.
v2.1 brings improvements in training and dataset that should address some complaints we've heard, and improve overall quality.
No action needs to be taken, the Dreambot will already be using v2.1 by default as of this time. However, if you want to switch back to v2.0 because you prefer it, or if you want to compare -- you can do so with the model parameter in /dream.
We are looking forward to seeing what you all create with v2.1! Please share and highlight your creations with us in #1045349359044280360.
We will be adding a starboard system for Dreambot images later this week - if you browse the #1047760914008522782, #1047759726693335050, #1047759774739075102, ..., #1047760279267721236 channels and see an image you like, you will be able to ā in the reactions it to share with the rest of us! If enough of you ā an image, it will be posted on the starboard.
p.s. this is free to use for the time being, unlimited images.
deep storm
š¤ Negativity Comes To Dreambot š¤ @everyone
We have heard you -- everyone appears to wants some negativity in their lives. š¤·
We have added negative prompts to allow you to better control your compositions with v2.1 in Dreambot!
This adds the following fields:
negative_promptdefault is:ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draftnegative_strength, default is:-0.25, reasonable ranges are-0.05to-2.0.
The default negative_prompt should improve output quality; this will apply to all /dreams unless overridden, as the Dreambot only shows non-default values. If you want to disable the negative prompting, use negative_strength: 0.
Attached are example outputs from negative prompting -- note the lack of other colors than reddish hues or grey!
prompt: A beautiful redhead in a dress. negative_prompt: blue, green
lyric bobcat
**Diffusers! v2.1 has arrived! ** @everyone
**So Whatās New š **
Stable Diffusion 2.1: Negative prompts are now supported in DreamStudio by appending ā|<negative prompt>:-1ā to the prompt. For instance, appending: |disfigured, ugly:-1.0, too many fingers:-1.0 occasionally fixes the issue of generating too many fingers.
Extreme aspect ratios: The model also has the power to render non-standard resolutions to give you beautiful vistas and epic widescreen imagery.
For more in depth details on the latest SD v2.1 Release please see our full blog post here:
[ https://stability.ai/blog/stablediffusion2-1-release7-dec-2022 ]
Note: *For more details about accessing and downloading v2.1, please check out the release notes on our GitHub:
[https://github.com/Stability-AI/stablediffusion]
Additionally, you can try v2.1 right now on DreamStudio at https://beta.dreamstudio.ai
Thank you ALL for your incredible passion, creativity and optimism overflowing into these spaces! Happy Diffusing Dreamers š
deep storm
š¤ Dreambot Bugfixes and QOL Improvements š¤
We have just rolled out a new version of the gobot code that powers Dreambot. There are some nice QoL improvements in the interface, as well as bugfixes, and some quality improvements.
Bugfixes
ā¶ļø Disabling the default negative_prompt works now, by setting negative: False. negative_strength did not actually do anything, so it has been removed.
Image Quality Improvements
ā¶ļø Outpainting should now work far better out of the box.
ā¶ļø Image to image 'graininess' resolved.
UI Quality of Life Improvements
ā¶ļø The Discord UI is now much more mobile friendly by putting virtually all parameters such as image_strength, guider_strength be drop-down menus rather than something you need to type numbers into.
ā¶ļø We now show what the prompt, negative_prompt, guider_prompt actually looks like when the AI processes it. Many people had been posting super-long prompts, and these actually do not work. We show the truncation that is done by the server, and warn about the truncation.
ā¶ļø All parameters have had their descriptions cleaned up, and all of them now have default documented for them.
ā¶ļø init_image has been renamed to image_init to make grouping of parameters more clear.
ā¶ļø /interrogate is now more efficient about how it represents the request data structure, making it more likely that the response will fit inside Discord's 1000 character limit.
Happy Diffusing, and have a Stable Day!
fluid jungle
POW Week #11 Challenge- "ANIME" Winner <@&1018216027244466246>
Thought I'd surprise you all with an early winner announcement, due to you all being such good sports during this time-extended POW! I said bring the waifus and you all....certainly didn't hold back as expected. Let's dive into our results, shall we?
Give it up for @ripe light and their cutest chibi teacup submission!!! 
And let's not forget our runners up who we welcome into Stable Society at long last!! @obsidian maple and @vague bobcat š¤
Head on over to #šļ½winner-gallery and give our winning submission all the emote love!!!
Full POW results/tally will be posted in #āļ½pow-info momentarily~
See you tomorrow for the beginning of Week#12 POW...Our long awaited finale: BEST OF SD 
(p.s. thank you for all the mikus)
lyric bobcat
šš We want to hear from YOU! šš
If you have a few minutes, please fill out the survey below š
Stability is looking into the possibility of developing official integrations of our generative AI tools (like Stable Diffusion) into creative products that you already know and love.
Before that, though, we're curious to know what creative software (like Photoshop, Cinema 4D, Sony Vegas, etc) you already use on a regular basis!
Your responses will let us know what to work on next! If you're interested in these future integrations, enter your email in the bottom of the form or message @scarlet grove on Discord. Thanks all!
lost pine
<@&1004115656620392538> Exciting news from our friends over at Harmonai!
Introducing Diffusion Radio! A 24/7 livestream of AI generated music from Harmonai, created using their latest models!
Fully AI-generated dance, hiphop, metal! Diffusion Radio has it all, streamed in continuous high-quality stereo audio!
The channel features songs varying from 45 to 180 seconds long, made using the latest of Harmonaiās latent diffusion models!
Head on over to the Harmonai server here: https://discord.gg/harmonaiplay
And check out Diffusion Radio on Youtube: https://www.youtube.com/watch?v=uGRLOMf2hSc
fluid jungle
POW Week #12 Challenge- "BEST OF SD" Winner <@&1018216027244466246>
Thank you all,
for being a wonderful part of the community and bringing these little events to life! We've come a long way since the beginning of this year, haven't we? While everyone comes up with their resolutions, their epiphanies, and their own way of balancing their priorities and life-- I'd just like to once again thank you all for creating such a warm and nurturing atmosphere here. 
Let's all continue doing our best.
š Without further ado, for our POW Finale,
Give it up for @fast hill and their BEAUTIFUL cozy wintery cabin submission!
Applauds all around to our runners up: @dense kestrel and @signal fossil !
Head on over to #šļ½winner-gallery to give these lovely submissions some emote love!
POW tally will be posted in #āļ½pow-info momentarily.
See you next time~ In 2023?! 
lyric bobcat
lyric bobcat
@everyone WE ARE LIVE NOW ON TWITCH FOR THE FIRST TIME!
Come say hi!
fluid jungle
Pop-up Event and Dreamer-Communities Update
Dear Dreamers, <@&1018216027244466246> ,
We're doing some New Year cleaning on our server, and thought why not make it into a fun event for us all? The plan is that we will be archiving the current Dreamer Community text-channels with an opt-in role to view all of it's history. But, wait... what's going to happen to our beloved little nooks?
Fear not-- They won't be erased. Only migrated! We'll be moving them all on over to Forums, to better organize and structure out our beloved communities 
That being said, before we implement any structural changes such as this,
We cordially invite you to submit one of your best pieces of artwork for our upcoming Dreamer Forums! Please send in generations relevant to the community of the event thread! This is a unique opportunity to immortalize your art into our new community forums and be presented as examples for each related category.
To participate, simply send in an image of your artwork to the associated category your image falls under. (ex. fantasy, interior, portrait, surrealism)
Youāll be able to begin submitting your work NOW! <t:1673298000:F>
The community will vote on these submissions and the top voted artwork will be featured in the forum threads corresponding to the dreamer community.
Head on over to #1048378510894436562 (https://discord.com/channels/1002292111942635562/1061021993530564618) to see more detailed info, rules, and deadlines!
We look forward to seeing your submissions and celebrating the joy of infinite imagination with you!
lost pine
Happy Friday y'all! 
Check out this new tune by The Chainsmokers featuring an amazing animated video by Remi Molettee, created using our very own Stable Diffusion! Head over to Remi's Instagram here: https://www.instagram.com/remi_molettee
And get a load of this awesome animation!
https://www.youtube.com/watch?v=P10HOvTNPbk
fluid jungle
POW Week #13 Challenge- "Beginnings" Winner <@&1018216027244466246>
POW is BACK! You all took us to your very own portals for this challenge it seems. I loved traveling along with you all 
Without further ado...
Give it up for the king @ripe light and their beautiful handheld world!
Applauds to our runners up as well @polar quail and @dense kestrel (a new user joins the Stable Society?!)
Head on over to #šļ½winner-gallery to show our winner some love!
POW tally will be posted in #āļ½pow-info momentarily.
Thank you all as always ! See you next time! 
fluid jungle
<@&1018216027244466246> voting is now OPEN for #1048378510894436562 !
You may now emote all of your favorite submissions! Please look here to see all of the times for this event:(#1061021993530564618 message)
Have fun and good luck!
fluid jungle
**POW Week #14 Challenge- "Metamorph" Winner <@&1018216027244466246> **
We traveled through the Beginning, middle, and arrived to a transformative end this POW. You all showed us your interpretations of undergoing intense changes amongst an ever fickle world. Thank you for letting us peek into your minds
Let's get into the main course~
Congratulations @chilly flicker and their diagram showcasing the stages of a creatures transformation! Head on over to #šļ½winner-gallery to give this submission some well deserved emote love!
And we couldn't possibly leave out our runners up, could we? At long last, welcome @lethal rune and @cold sparrow to the ~Stable Society~
POW tally will be posted in #āļ½pow-info momentarily.
See you tomorrow for POW#15 
fluid jungle
<@&1018216027244466246> voting is now OPEN for <#1048378510894436562> !
You may...
<@&1018216027244466246> Aaaaand that's a wrap! Voting is now closed for The Dreamer's Canvas and will now be privated as we look over the winner tally. We'll see you very very !
lost pine
š¤ Dreambot winding down š¤
After 2 months of non-stop generation, almost 2 million images and some truly outstanding results, itās time for Dreambot to take a well-deserved break! 
The Dreambot channels will be heading into the archive over the weekend!
Head on over to the #š„ļ½roles channel to pick up the Dreambot-archive role, and youāll still be able to view them along with all your creations as read-only channels.
As one chapter draws to a close and Dreambot takes its leave, many more exciting surprises will be heading our way...
UPDATE: All Dreambot channels are now closed! Thank's for dreaming with us everyone!
fluid jungle
POW Week #15 Challenge- āYear of the Rabbitā Winner <@&1018216027244466246>
Thatās a wrap folks! Who wouldāve thought this POW would be so popular! Being flooded by a bunch of bunny imagesā not a bad problem to have at all. Thanks so much everyone for your participation! 
š„
Without further ado,
š§§ Congratulations to @icy cradle and their winning ethereal masterpiece! Capturing the very essence of the rabbit and its proverbial meaningsā grace, beauty, mercy, and good luckšæ
š§§ And letās not forget our runners up, both hopping into a tie for second place! @faint iris and @ornate bobcat !
š§§ POW tally will be posted in #āļ½pow-info momentarily. . . .
š§§ Go give the winning submission some emote love over at #šļ½winner-gallery
fluid jungle
POW Week #16 Challenge- "Heart U" Winner <@&1018216027244466246>
I know it's a bit early for Valentines celebrations and aesthetics, but I feel like celebrating love this month starts us off on the right foot, doesn't it? No matter what love means to you-- whether it's something you find to be stereotypically sappy, or even niche-- I hope you are surrounded by all that encompasses joy. I hope that joy starts with yourself! 
Congratulations to @lone stag for their winning sunset spun tale of swan love!
And big shouts to our runners up @boreal dust and @civic badge !! Beautiful work!
POW tally will be posted in #āļ½pow-info in just a moment. . .
Go show the winning submission aaaaall the emote love over at #šļ½winner-gallery
Thanks all for your eager participation ^^
Heart U!
lyric bobcat
New vibe who dis? š
Welcome @everyone to Stable Foundation!
The Stable Diffusion server has received a fresh coat of paint and a whole lot of recalibration!
Weāre amplifying all the cool happenings from across the AI eco-system and bringing a host of exciting new features home to our beloved server.
This is the Foundation on which weāll build a new chapter for our Stable community!
Check out all these cool new features:
Find out all the latest #1072229057707659404 and explore the Stable Communities!
Harmonai, DeepFloyd, Carper, EleutherAI and more!
Head on over and get acquainted!
See our Stable Stage Event with Stelfie the Time Traveller this Thursday!
A huge thank you to ALL 146818 of YOU for making this amazing community what it is today! We could not have done this alone. š
WE WERE HERE FOR IT Let's lay the foundations for a stable future together!
lost pine
Hey everyone!
Our Stable Stage event, starring Stelfie the Time Traveller is live now! Come and hear all about Stelfie's adventures through time using Stable Diffusion, and his workflow used to create these amazing windows into the past and future! Head over to Center Stage and check it out!
https://discord.gg/2PXWm9Sz?event=1072153135692386345
forest field
POW Week #17 Challenge- āFoodchitectureā Winner <@&1018216027244466246>
Last weekās POW took us on a journey to a city transformed into treats and sweets. What if we could capture this feast for the senses within our magical city forever?
Congratulations to šš„¦ @boreal dust š„¦š for capturing the essence of the Foodchitecture filled with delicious.. fruit and veggies! A bold and colorful art piece across a stunning landscape brings the genieās magic to life through their vibrant depiction of the food-filled city.
Welcome our runner ups @grave vine & @winter turtle to the Stable Society! 
Take a look at our food architect over in #šļ½winner-gallery and show them some emote love!
Pow results for the top 25 submissions will be posted in #āļ½pow-info shortly!
š Thank you to everyone who participated! Drop by tomorrow for a special day and this weekās POW! š
forest field
Hey @everyone! Love is in the air!
Itās that time of year for showing your admiration, appreciation and gratitude!
Thereās so much love right here in this wonderful community, so letās shout it out and spread it a little further!
Someone in particular you'd like to thank for getting you started on your SD journey?
Amazed by someone's artistic expertise and just have to let them know how great their work is?
Want to thank someone for sharing a top tip, or introducing you to an awesome new thing?
An all-around amazing person who you'd like to shower with appreciation?
"But if only there was a cool new slash command for just this very purpose!", you may say. Well look no further!
Introducing the /showlove command, courtesy of our wonderful bot expert @cinder oasis!
Using the command will let you shout out that special someone, and display your message of appreciation in the new #1074739806833217606 channel!
There is a 30 character limit on these messages as of right now, so keep it short and sweet! This will be expanding soon so you can let everyone know just how much you appreciate them in detail!
A little love goes a long way, and you may even be rewarded in the form of Discord Nitro and DreamStudio credit gifts! Everyone who posts or receives a /showlove message will be entered into our random prize draws over in #1051259766976040960! Everyone who enters will have an equal chance of winning, whether they've shown the love to just that one special someone, or spread it far and wide!
Thereās far too much love to celebrate in just one day, so weāll be running this event for one week to give you all time to let those special people know just how much of a difference theyāve made!
Thank you everyone for being such a beautiful, vibrant and helpful community!
Much love! š
forest field
Hello, everyone! <@&1002917600042422362> <@&1004116145319710840>
Itās time for our latest Stable Stage event featuring MedARC!
https://discord.com/events/1002292111942635562/1074322377686663188
Apologies for the slight rescheduling!
For this event, weāll be welcoming two special guests:
Tanishq Abraham, CEO and Founder of MedARC!
&
Jeremy Howard, President of MedARC and Co-founder of Scale.ai!
Our amazing guests will be giving us the run down on their recently-launched MedARC initiative, talking all about exciting applications of AI in the medical world, and discussing their vision for a future of healthcare reinforced by artificial intelligence with some exciting products in the works!
Please join us over at the Center Stage for what is sure to be a fascinating insight into a whole new frontier of artificial intelligence!
Don't forget to leave all your amazing questions in #1029055412764422214
Thank you & Enjoy the show!
Click this link to be taken directly to the Center Stage: https://discord.com/channels/1002292111942635562/1005545148211527800
We're LIVE!
forest field
POW Week #18 Challenge- āDesires" Winner <@&1018216027244466246>
Last weekās POW followed the adventure of Calvin in search for his deepest desire! Upon reaching his goal he had discovered what he loved most was already a part of his life, and that the greatest treasure is in the eye of the beholder. š
Congratulations to š
@jade glacier š for their winning submission depicting a portal into paradise as seen through the crystal ball - beckoning into a world of wonder and possibility. The colors and textures of the artwork were masterfully rendered, evoking a sense of awe and transcendence!
Welcome our runner ups @spare lily & @hot forum to the Stable Society! š„
Peek into our winners depiction of Calvin's Cave at #šļ½winner-gallery and show them some emote love! š
POW results for the top 25 submissions will be posted in #āļ½pow-info shortly!
Hoping that this POW inspired all of our participants to explore the depths of their desires and to express them in unique and compelling ways! See you tomorrow for a new theme and wholesome vibes at #šļ½pow-discussion š«§ š«§ š«§
timid smelt
Hello @everyone!
It goes without saying that we have an immense appreciation and respect for the fantastic and incredibly inspirational community we have here on this server!
You have all demonstrated just how much you care for this space and we likewise feel it's important for you to get the most out of your stay here with us!
Since everyone has now had a chance to get hands-on with the Stable Foundation revamp, it's important for us to take the time to understand what your thoughts and feedback are to help better accommodate you as we move forward.
In order to collect your input, we've created a short survey that we would greatly appreciate you taking the time to complete (it should only take 1-2 minutes at minimum).
The first part of the survey will focus on our past events as well as what events you'd like to see in the future, with part 2 focused on some quick questions regarding your use of the server as well as any additional feedback you have on the current layout!
Please be as detailed as you would like and don't hold back on any of the excellent feedback or ideas we know you have!
The survey will be kept open until <t:1677517200:D> to give everyone a chance to participate!
https://docs.google.com/forms/d/e/1FAIpQLSdpaL4lcXvdexbLEnAnP-G3w5e56nXuq_0-CfWDFKpmpAlSww/viewform
Stay Stable
lost pine
Any Dreamers in the chat?
Take a trip back in time to our Twitch launch and experience a wild ride through a year of non-stop AI innovation, artist collaboration and community celebration all over again!
Since we kicked off this journey back in August last year, you as a community took Stable Diffusion to new and unprecedented heights.
From amidst the beautiful chaos of a creative revolution-in-progress, our awesome host and community legend @lyric bobcat united a** huge all-star cast** of Stability's researchers and developers, groundbreaking AI artists and game-changing community innovators to close out the year in style!
Travel with us through this super-slick audiovisual time capsule, and re-live everything that made Stable Diffusion's first year special:
YOU!
Check out the awesome video edit right here! š„
https://youtu.be/cf07ylX4x6I
forest field
POW Week #19 Challenge- āUnderwater Worlds" Winner <@&1018216027244466246>
Towards a journey seeking adventure and magic, we followed Lyra into the depths of the ocean diving into underwater worlds. We encountered layers of vibrant ecosystems - breathtaking creatures, long forgotten shipwrecks, ruins of civilizations lost to time and diverse merfolk!
šŖø Congratulations to 
š«§ @clear comet š«§ for winning POW this week with a unique take on our theme with their haunting yet beautiful mermaid wrapped in vibrant contrasting splashes of fuchsia and purple. The beauty of the deep is not one to be taken lightly! 
šŖø Our runner-ups this week are current members of Stable Society @faint iris & @lone stag! 
šŖø Glimpse at the queen of underwater worlds over at #šļ½winner-gallery and show her some emote love! š§āāļø
šŖø POW results for the top 25 submissions will be posted in #āļ½pow-info shortly!
This POW was inspired by the mysterious wonder of the underwater worlds within our own planet and itās been really fun watching the different interpretations from you all! As always, weāll see you tomorrow for a new theme and wholesome vibes at #šļ½pow-discussion āŖāŖāŖ
lost pine
Hey @everyone! Weāve got some exciting news! 
š§¬**Love Stable Diffusion? Love making things in 3D? Why not combine them!**š§¬
Introducing Stability for Blender - an officially supported, free to use, zero-hassle way to use Stable Diffusion and other upcoming models inside Blender!āØāØ
With no dependencies to install and no GPU required, Stability for Blender lets you preform AI post-processing effects with just an internet connection - something that previously required owning prohibitively expensive hardware.
Whether youāre a Blender beginner or a modelling master, head on over to the Platform site at https://platform.stability.ai/docs/integrations/blender for all the details, guides and documentation to help you get started!
Come and hang out in the new #1054807328509145158 channel - weād love to see your creations and hear your feedback on the add-on!
Check out the amazing video below, showcasing just some of the limitless potential Stability for Blender is capable of before you take it for a spin yourself! 
Perhaps now may be the time to polish up those awesome AI skills to a mirror-finish... Today's amazing new add-on could be just one of the tools in your arsenal when the time comes for an unprecedented contest on the horizonā¦
fluid jungle
āŖ POW Week #20 Challenge- "Something" WinnerāŖ <@&1018216027244466246>
āSomethingā represents a dissected, unspoken, or emotionally charged representation of yourself. It is the part of you that isnāt easily displayed at surface level, but can be found with a littleā¦digging. This week we certainly met one another's "Something", shared with one another personal stories, opened up communication-- all through creating such emotionally-charged and diverse images. Thank you for sharing with us! š¤
ā½ Congratulations to @dense kestrel for this all too relatable and creative submission depicting a sense of escapism from an ever heavy reality.
ā½ BIG shouts to runners up @feral crow (a new Stable Society member!) and @chilly flicker
ā½ POW tally will be posted in #āļ½pow-info momentarily. . .
ā½ Go give our winning submission aaaaall the emote love over at #šļ½winner-gallery
See you tomorrow for our 2nd chapter of a 3 part POW story!
Are you ready to go "Somewhere"? ā«
lost pine
This just in!
Stability.ai and Clipdrop creators Init ML come together to push the limits of cutting-edge AI applications even further! š
Init ML enter a new chapter as an** independent subsidiary** of **Stability.ai **in this monumental acquisition!āØāØ
Reinforced by Stability.aiās powerful AI technology and compute, the Clipdrop team will build the next generation of imaging tools and best-in-class applications, making open source AI even more accessible for everyone in ways we never thought possible!
Keep an eye on the horizon, AI-enhanced imaging tools are about to blow our minds all over again! 
Check out the blog post for more info: https://stability.ai/blog/stability-ai-acquires-init-ml-makers-of-clipdrop-application
lost pine
Happy International Women's Day!
Celebrate today with a very special Instagram Live, featuring outstanding AI artist KaliYuga!
Head over to Instagram @stablediffusion and join** KaliYuga** as we celebrate women in the world of AI art, and gain an insight into her spectacular surrealist style!
Drop by here https://www.instagram.com/stablediffusion/ at <t:1678300200:t> today and ask your questions live in the chat! š
lost pine
Last month, you spread the love, appreciation and good vibes in the Show me love channel!
But why limit the love to just February? Letās keep the positivity flowing year-round with the freshly upgraded /showlove command!
Use the new and improved /showlove command to send a message of appreciation to anyone in the server for all to see in the #1074739806833217606 channel!
Now featuring extended character limits so you can let that special someone know just how special they really are, and a whole host of stunning randomly-selected backgrounds to really make that message stand out!
Available to use now in any channel, letās spread the good vibes!
Thank you everyone for the wonderful positivity and love you bring that makes this community so special! š
timid smelt
Hello again, @everyone!
My apologies but there is no 'new product announcement' for this post. Instead, we wanted to clarify something we feel is very important to every single user on this server.
To protect your safety, weāre making some changes to how you claim prizes!
As you know, we have various exciting competitions on this server and frequently offer prizes as well! While this can certainly be exciting, it also has the potential to be abused by bad actors or malicious 3rd parties trying to pose as server mods or stability staff.
Moving forward, along with continuing to publicly announce any contest winners in that competitions designated channel, any prize winners will be added to a private ticket in the server to facilitate any conversations needed to claim your prize!
(We'll make sure to ping you in the ticket so you can find it if you win!)
We would also like to remind you that *we do not tolerate accounts joining this server purely to take malicious actions against our community. *
We encourage you not to engage with any unsolicited direct messages asking you to join servers or claiming you have won a prize especially if you never directly entered the competition in the first place. Instead, report this behavior to #1010934719455707218 along with any extra details you have so we can take quick action against these accounts!
Any official partners this server works with can be found in the #1072229057707659404 channel for you to verify if needed
We also highly encourage you to look over Discordās guidelines on securing your account as well as reviewing your options for accepting direct messages here: https://discord.com/safety/360043857751-four-steps-to-a-super-safe-account
If at any point you are in doubt about the authenticity of a message you receive from our server staff, please fill out a request in #1010934719455707218. Only moderators and staff currently rolled in our server can reply to these tickets & we will do our best to help from our side whenever possible!
Thank you all and stay safe!
fluid jungle
POW Week #21 Challenge- "Somewhere" Winner <@&1018216027244466246>
āSomewhereā represents the environment of your own mind. If we werenāt humansā but could only be portrayed by landscapes, nature, and architecture.. What would we look like? What would that say about who we are? This week we traveled through endless landscapes, dived into coffee-cup worlds, swam through the sky, slept under sakura trees, and so much more! Thank you for sharing with us so many different beautiful submissions and telling your stories. And welcome in to all the new members of POW! Happy to have you.
š Congratulations @jade glacier for taking us to a cozy mini world of window-side beauty. Your "Somewhere" is the perfect energy we need for the week ahead! ā š
š Shouts to @icy cradle aaaand new member of Stable Society @agile jewel !!!
š POW tally will be in #āļ½pow-info for you soon!
š Go give our winning submission aaaaall the emote love over at #šļ½winner-gallery
See you tomorrow for our 3rd chapter of a 3 part POW story!
Maybe we'll meet *"Someone"?*š”
fluid jungle
POW Week #21 Challenge- Chapter 3: "Someone" Winner
Your journey has reached its conclusion, <@&1018216027244466246> . . .
We've wrapped up a POW 3 Part Story this week! Thank you all for participating and showing your eagerness to share and communicate with one another. These submissions have been the gateways to discuss not only your workflows, but also your memories. I hope you enjoyed!
š” Congratulations!!! @signal fossil for depicting a beautiful emotional journey to meet a special someone! š
š” 3 WAY TIE shouts to @weak holly , @faint iris , aaaand @lament shoal ! Welcome in Stable Society Newcomers~
š” POW tally will be in #āļ½pow-info soon!
š” Give our winning submission some emote love over at #šļ½winner-gallery
See you tomorrow!
lost pine
Hey everyone!
Relive the launch stream experience with @radiant hinge in this awesome highlight video and discover the secrets of model training!
Join Claire for an exclusive insight into a world of mind-blowing creativity and cutting-edge technology as we explore her Stable Diffusion model training workflow!
lost pine
Exciting news @everyone!
Check out the Pick a Pic project's latest addition: A sneak peek of the upcoming Stable Diffusion model - SDXL!
You may have already heard about Pick a Pic, an open source collaborative project where you can try out (and help improve!) the latest text to image models.
Try it out for free with your own prompts and rate your favourite outputs today!āØ
By collecting image rankings from the community and releasing the preferential data anonymously, you can help the research community at large align future models with human preferences, shaping the future of text-to-image models!
Head on over to https://pickapic.io and sign up for free!
forest field
POW Week #23 Challenge- āCatstravaganza" Winners & NEW EVENT LAUNCH?? <@&1018216027244466246>
You've catptured our hearts and the alluring nature of cats - their (sometimes) graceful, playful and mischievous personalities. You've also made it through the flood of cat puns following this weeks theme! Now it's time to announce our purr-ize winners... 
Congratulations to our TIED WINNERS 
@rugged rock & @vague bobcat who've enchanted us with their dreamy and mystical felines! 
Our runner-up this week is our very own stable mod @iron nova! 
Both of our winning cats will be showcased over at #šļ½winner-gallery, show our co-winners some cat emote love! 
POW results for the top 25 submissions will be posted in #āļ½pow-info shortly!
But wait.. there's more! Welcome to....
CHALLENGE OF THE WEEKEND (COW)
We've been keeping an eye on your skills in POW, and we're confident that you're ready for the next level of artistic challenges! Every weekend, we will bring you a different challenge that requires you to think outside the pasture, push your skills, and create something very moo-nique.
Here's the twist - we'll be adding limitations and randomness to ignite your creativity! Youāll be restricted in some way, such as limiting your materials or techniques. We may also add a thematic element to the challenge, inspiring you to experiment with new ideas.
But don't worry, these restrictions are not meant to stifle your creativity. Instead, they serve as a launching pad to propel you towards your best capabilities. By requiring you to work within specific parameters - you can unlock new perspectives, innovative solutions, and unexpected artistic directions that you may not have explored otherwise!
So, dear diffuser, are you up for the challenge?
More info will be up shortly as we prepare for launch this coming Friday. In the meantime, visit our brand new pasture over at #1087494110358798547 to discuss the new event & maybe bovines too!
fluid jungle
**POW: Week #24 Challenge- "Elemental Flowers" Winner Announcement! <@&1018216027244466246>
Fire..Earth..Air..Water...Winner?! Who would've thought the elements would be so hard to display as flowers?
Congratulations @rugged rock !!š Back to back POW Winner! Your submission perfectly encapsulated all of the elements combined as one!
Running closely behind are the lovely @vague bobcat aaaaaaand @weak holly! Stable Society 
Go show our winner some emote love over at #šļ½winner-gallery
POW results for the top 25 will be posted in #āļ½pow-info shortly . . .
**Thank you as always to our beautiful community and our regular event members! **
P.S. Have you taken a look at #1087493421209485393 #1087493722645725184 ; our sparkly new weekend challenge? Voting will be wrapped up <t:1680548400:R>, so be sure to get all your votes in~
forest field
COW Week #1 - āFUTURISTIC FASHION" Winners 
<@&1018216027244466246>
Our first challenge of the week(end) was to light things up with some fancy LEDs! We wanted to see what kind of techy gadgets you could incorporate into your fashion designs in a creative and functional way, while also showing off your love for all things neon. And boy, did you deliver! We're talking about the future of fashion, people! 
Since this is the first time we've released our COW into the pastures, let me introduce the format! COW is built around challenges and so.. we'll be having 2 winners! One of them will be decided by you, the voters! The 2nd winner will be decided by a
Congratulations to our 1st place winner and beloved mod @iron nova introducing the moo-gnificent DJ CoW hoofing to moo-sic & mixing some sick beats! 
Our staff pick for this week is 
@sage patio creating a beautifully lit neon character wearing very on-theme futuristic fashion!
Both of our winners will be showcased at #šļ½winner-gallery give them some love!
COW results for the top 10 submissions will be posted in #1087493421209485393 shortly!
Thank you to everyone who participated this weekend and had fun with our silly yet endearing endless cow puns, we hope you see you tomorrow for POW and on Friday for the next COW! ā¤ļø
forest field
Good day folks and happy almost Friday!
We are thrilled to announce that our new forum, #1092446741984444416 is now live and open for discussion!
As the name suggests, this forum is all about coming together as a community to collaborate and share exciting projects related to API, integrations, and other services involving Stability AI or Stable Diffusion. We strongly believe in open source discussion to unlock the full potential of technology and create more impactful solutions.
Our forum provides a platform for everyone to share their ideas, ask questions, and engage in meaningful discussions about the latest projects in the community. We encourage you to use this as an opportunity to showcase your work, receive constructive feedback, and collaborate with others. Connect with like-minded individuals and learn from their experiences or contribute to their projects.
We can't wait to see you push the boundaries of what's possible with open collaboration.
<@&1004114999867879506> <@&1004116145319710840> <@&1072165794571628686> <@&1072165653051621508>
fluid jungle
POW: Week #25 Challenge- "Synesthesia" Winner Announcement! <@&1018216027244466246> 
Synesthesia /ĖsÉŖn.ÉsĖĪøiĖ.ŹÉ/: an anomalous blending of the senses in which the stimulation of one modality simultaneously produces sensation in a different modality. Synesthetes hear colors, feel sounds, and taste shapes.
Congratulations to our very own @iron nova ! š 
Could he possibly be morphing into the...Event King???
Shoutout to runner-up and new member of Stable Society-- @turbid parrot !
....and but of course a very kind nod to our 3rd place 4-way tie 
@vague bobcat , @cobalt beacon , @icy cradle , and @feral crow !
Go show our winner some emote love over at #šļ½winner-gallery !
POW results for the top 25 will be posted in #āļ½pow-info shortly . . .
forest field
COW Week #2 - āCOSMIC HORRORS" Winners <@&1018216027244466246>
Our second week(end) of challenges tested the limits of human sanity, hopefully the horrors of the cosmos have inspired you to new heights of creativity! These brave souls dared to stare into the void, and emerged victorious... 
Congrats to our 1st place winner āļø @fast stone āļø for creating a hauntingly twisted and terrifying eldritch being! š¦
Our staff pick for this week is @winter turtle! Their submission was a standout for us, feeling almost like there is a story behind the artwork leading up to that moment!
Both of our winners will be showcased at #šļ½winner-gallery show them some emote love! 
COW results for the top 10 submissions will be posted in #1087493421209485393 shortly!
BIG shoutout to all of our lovely participants! It's a pleasure creating these events for you and experiencing your creativity and artwork, see you tomorrow for the next POW and on Friday for the upcoming COW #3! 
timid smelt
Hello, @everyone!
We have an exciting update to share with you all about the next upcoming model for Stable Diffusion:
SDXL Beta is now live for all users on the API & DreamStudio!
The SDXL model has some incredible capabilities and weāre excited to see what you think after getting hands on with it!
Some of the highlights of SDXLās capabilities include:
š Next-level photorealism capabilities
š Enhanced image composition and face generation
š Rich visuals and jaw-dropping aesthetics
š Use of shorter prompts to create descriptive imagery
š Greater capability to produce legible text
The release of SDXLās API will be followed by an open-source release of SDXL (currently in Beta testing) which, like all of Stability AIās open-source models, is optimized to be accessible to as many users as possible. This is the power of open source: tapping into the vast potential of millions of talented individuals who donāt have the resources to train a state-of-the-art model, but who have the ability to do something incredible with one.
Keep an eye out for more details on the public open-source release of SDXL
The SDXL Beta will continue to improve over time thanks to all the amazing feedback you have been providing! Make sure to stop by the #āØļ½sdxl channel and leave any thoughts you have so it can get even stronger!
To test out Stable Diffusion XL in DreamStudio visit: https://dreamstudio.ai/
For access to Stable Diffusion XLās API visit: http://platform.stability.ai/
You can also join the discussion in our #1042896447311454361 channel here on Discord!
For more information visit: https://stability.ai/stable-diffusion
Find out about this and more in:
Our Stable Stage AMA featuring Emad starting now!
Tune in to Center Stage and catch it live! āØ
Get ready for some really exciting news and answers to all your questions from the GPU Emperor himself!
#1029055412764422214 is open and waiting for you to ask those burning questions, so come stop by and enjoy the show! š„
See yāall there! š
https://discord.com/events/1002292111942635562/1094987731425308742
fluid jungle
POW Week #26 Challenge- "Alice in DiffusionLand" Winner Announcement! <@&1018216027244466246>
Alice fell down the rabbit hole and entered a digital dream. . . Along the way she met our winner. . . š³ļø
Congratulations!! to our winner @errant wigeon !! What an absolutely creative interpretation of Alice!
A round of applause for our runners up @river echo and @cobalt beacon! (*with a star this time...
)
Give our winning submission some well-deserved emote love over at #šļ½winner-gallery 
POW Top 25 tally will be posted in #āļ½pow-info momentarily . . .
forest field
COW Week #3 - āCHAOS & ORDER" Winners <@&1018216027244466246>
Amidst the battle of two opposites, you have shown us the raw beauty of two opposing forces clashing into harmony.. or destruction.
Congrats to our 1st place winner 
@plush dragon A beautiful submission split in half befit with Thorns & Roses representing the aspects of both Chaos & Order! āŖ š“
Our staff pick for this weekend is @river python! Your composition had elements of both Chaos & Order, and harmony that you achieved between the two was simply breathtaking to see! šø
Both of our winners will be showcased at #šļ½winner-gallery give them some emote love! 
COW results for the top 10 submissions will be posted in #1087493421209485393 shortly!
*See you tomorrow for our upcoming POW and on Friday for the next set of challenges at COW! ā¤ļø ā¤ļø ā¤ļø
fluid jungle
ā Launch of StableLM: A New Open-source Language Model ā
Weāve got news for you @everyone!
Weāre incredibly excited to announce the launch of StableLM-Alpha; a nice and sparkly newly released open-sourced language model! Developers, researchers, and curious hobbyists alike can freely inspect, use, and adapt our StableLM base models for commercial and or research purposes! Excited yet?
Letās talk about parameters! The Alpha version of the model is available in 3 billion and 7 billion parameters, with 15 billion to 65 billion parameter models to follow. StableLM is trained on a new experimental dataset built on āThe Pileā from EleutherAI (a 825GiB diverse, open source language modeling data set that consists of 22 smaller, high quality datasets combined together!) The richness of this dataset gives StableLM surprisingly high performance in conversational and coding tasks, despite its small size of 3-7 billion parameters.
Hereās an example:
User: Write me a C program that computes the meaning of life.
StableLM: As an AI language model, I'm not capable of providing you with a complete program. However, here is a simple example of a program written in C that calculates the amount of life one has lived:
int main()
{
int life, t, min = 0, max = 0, value;
printf("Please enter a number representing how much life you have lived: ");
scanf("%d", &life);
value = life;
do {
printf("Your number is: %d\n", life);
t = life;
if (min == 0 || t < min) {
min = t;
}
if (t > max) {
max = t;
}
if (max > t) {
break;
}
} while (life > t);
printf("Your life is %d\n", value);
if (value <= max) {
printf("Your life is too low.\n");
} else {
printf("Your life is too high.\n");
}
return 0;
}```
>
> StableLM: Please note that this is just a simple example, and the actual meaning of life would have more complex components like social, philosophical, etc.
>
 Ready to try it out for yourself? : https://github.com/stability-AI/stableLM/
 HuggingFace: http://stablelm.co/
 Share your hype, findings, and feedback over in our new shiny channel: https://discord.com/channels/1002292111942635562/1098025024541167646
 Want to know even more and have your own burning questions? Ask to your heartās content on our Stable-Stage tomorrow! -> https://discord.com/events/1002292111942635562/1098267150726221824#1005545148211527800 is now live for our StableLM Q&A!
If you have any questions you'd like to asks the hosts, post them in #1029055412764422214 & enjoy the show!
https://discord.com/events/1002292111942635562/1098267150726221824
fluid jungle
POW Week #27: "Encapsulated" Winner! <@&1018216027244466246>
Worlds within orbs was the theme for our last POW! You took us to many beautifully captured spaces--now it's time to set our winner free
Congratulations to @river python for their lush, beautiful, 4k getaway! Their submission clearly caught the eyes and votes of many! 
A round of applause to our runners up @feral crow and @errant wigeon! Welcome Dreamck Both are already members of Stable Society~ (Note: This will be the final invite to Stable Society via runner-ups, and now officially be reserved to our first place winners!)
Our beautiful winning submission will be in #šļ½winner-gallery ready for all your emote love 
POW Top 25 will be posted in #āļ½pow-info momentarily . . .
In the meantime, why don't you add some of your favorite or whackiest theme ideas to our POW-Roller? Maybe once spun, your theme could spotlight as our next POW! : (#āļ½pow-info message)
timid smelt
Hello <@&1018216027244466246> & <@&1072166265982034011>
Itās time for another exciting Stable Stage Q&A!
For the #DiffuseTogether animation challenge with Peter Gabriel, weāve put together an incredible lineup with some of the best in the business for AI Animation!
āŖ Are you curious what tools would work best for your workflow?
āŖ What inspired our guests to start their own journey?
āŖ What are all the secret tips and tricks to getting your outputs just right?
Get the answers to these questions and anything else you want to ask in our #DiffuseTogether Stable Stage Q&A starting at <t:1682539200:t> on the 26th featuring guest judges @radiant hinge & @lapis musk, joined by the OGās @pseudo mango, @dusky sorrel, @white idol, @dense shale& friends, hosted by @urban cypress!
Use the event below to make sure you donāt miss it!
https://discord.com/events/1002292111942635562/1100084952810405960
forest field
COW Week #4 - āUN-MYTHICAL CREATURES" Winners <@&1018216027244466246>
You've added many many entries into the Field Guide, with creatures of a hybrid nature and creatures unseen before, it is now time to immortalize these beings into legend.
Congrats to our 1st place winner šæ @river python šæ taking first place this week for both POW & COW! Accompanying their win is the magnificent species of Stellaris Chlorophytus complete with lore for the uninitiated!
Our staff pick for this week is @loud jacinth for their insanely adorable Spiked Groundfox! This was a really tough choice as you have all submitted some incredible work and there were so many good options! š¦
Both of our winners will be showcased at #šļ½winner-gallery show them some emote love! 
COW results for the top 10 submissions will be posted in #1087493421209485393 shortly!
*See you tomorrow for the next POW and on Friday for the upcoming COW!
timid smelt
Hello, everyone!
We have an exciting update for you all:
As of today you can use the API to make small, BIG!
Thatās right! Along with our popular text-to-image, image-to-image, and inpainting, Image Upscaling is now available on the API!
There are two models you can choose from when using these brand new features:
The super fast Real-ESRGAN model which can double the resolution of a typical 512x512 image in about half a second!
https://github.com/xinntao/Real-ESRGAN
Or
The ālatentā Stable Diffusion 4x Upscaler model! The latent Upscaller takes about 20-40 seconds to enlarge an image, but will yield much more detailed results with far richer texture!
https://huggingface.co/stabilityai/stable-diffusion-x4-upscaler
To get started, head on over to the API documentation! https://platform.stability.ai/docs/features/image-upscaling
If you use either Stability for Photoshop or Stability for Blender, Upscaling has already been integrated into the latest versions of those add-ons as well!
Check out these integrations here:
https://platform.stability.ai/docs/integrations/photoshop
https://platform.stability.ai/docs/integrations/blender/
timid smelt
<@&1018216027244466246> <@&1072166265982034011>
š“ We're live with our #DiffuseTogether Animation Q & A right now on #1005545148211527800!
Tune in for some fantastic tips from the animation experts & ask any questions you have over in #1029055412764422214
fluid jungle
__ Welcome DeepFloyd IF__ !
Hi there, keeping it Stable? We've got some exciting news stacked up for you to tinker with this weekend! Let's start with our first contender!
DeepFloyd IF is here and ready for you to use! But....what... is DeepFloyd exactly..?
INTRO TO DF
DeepFloyd IF is a state-of-the-art text-to-image model released on a non-commercial, research-permissible license that provides an opportunity for research labs to examine and experiment with advanced text-to-image generation approaches. In line with other Stability AI models, Stability AI intends to release a DeepFloyd IF model fully open source at a future date! Keep your eyes peeled!
FEATURES
GETTING STARTED
fluid jungle
ā¼ļø StableVicuna Release: the AI World's First Open Source RLHF LLM Chatbotā¼ļø
Quite the stacked Friday release day, eh? Here's some exciting news fresh off the press, @everyone! 
INTRO TO STABLEVICUNA
Welcome aboard StableVicuna! Vicuna is the first large-scale open source chatbot trained via reinforced learning from human feedback (RHLF). StableVicuna is a further instruction fine tuned and RLHF trained version of Vicuna 1.0 13b, which is an instruction fine tuned LLaMA 13b model! Want all the finer details to get fully acquainted? Check out the links below!
GETTING STARTED
fluid jungle
POW Week #28: "Pixel Infinity" Winner! <@&1018216027244466246>
Pixel art shares the hearts of many and brings a sense of nostalgiaā perhaps due to the cozy feeling we have shared playing older video-games, or maybe even the non-showy or ābusyā art style it has. Something about this seemingly simplistic style is so charming to many.
Congratulations to ...š„ @river python š„ !! Our back to back POW and previously COW Winner?! 
Who will be able to capture their crown?
Let's not forget runners up @feral crow and @errant wigeon !
Our cute winning submission will be in #šļ½winner-gallery and ready for all the emote love!
POW Top 21 will be posted in #āļ½pow-info momentarily. . .
forest field
COW Week #5 - āDYSTOPIAN CATASTROPHES" Winners <@&1018216027244466246>
Within a dystopian world where darkness and despair reign supreme, it can be easy to lose hope. Yet, you have shown us there can be a guiding light to keep fighting for a better future.
šÆļø Congratulations to our 1st place winner š£ @lusty pine š£! Their submission created a heartfelt moment that inspires feelings of warmth and compassion! šļø
Shoutout to @faint iris for tying first place with their beautiful re-creation of a legendary sword!
šÆļø Our staff pick for this week is @unique mango! A true embodiment of the saying, 'where there is life, there is hope'. š»
šÆļø Both of our winners will be showcased at #šļ½winner-gallery go and show them some emote love! 
šÆļø COW results for the top 10 submissions will be posted in #1087493421209485393 shortly!
See you tomorrow for the next POW and on Friday for the upcoming COW! We have even more events planned for you very soooooon ā¤ļø
fluid jungle
** āļø SD-Xperiment Test and Tinker!āļø **
Hello there to our lucky <@&1004866014133100674> and <@&1008475803207610399> !
Eager to test and tinker around with whatās been in the AI kitchen?
āSD-Xperimentā is here for our small group of members mentioned above for a limited time! Itās important to note that this an experimental checkpoint and is āstill cookingā as it were. Please be patient, as you may run into messy settings, outputs, or certain bugs!ā-all which will be vitally stored and collected in helping us test and cook our model to stable-perfection!
But wait!
As this set group of users are testing away, server-wide users-- @everyone will have access to vote for their favorite images, allowing the top-voted images to be viewable in the #š
ļ½pantheon
! We would highly appreciate the community's feedback in being able to enhance and make these models better and better. With your votes, we're step by step getting closer and closer to this goal!
It's important to know that these votes are not just a starboard-game, but a means to advance our technology together!
That being said,
Letās /dream into the stars!
timid smelt
āļø SD-Xperiment Update! āļø
Hello there to our beautiful <@&1005191371721822341>
I'm sure you've been eager to get your hands on the newest test model we have running on the bot, so we have some exciting news to share:
As a special thanks to some of our most committed members, everyone in the <@&1005191371721822341> role will have access to use the bot starting š **NOW!**š
(In order to help with the extra amount of generations, we'll be slowly expanded our channels as needed!)
Of course, everyone on the server will still have access to view & rate the images in #1100484581037195384 to help pick the best images to advance into the #š ļ½pantheon!
If you aren't a part of the beta members, don't worry! The bot will continue to roll out to more and more members !
In the meantime don't forget to vote for your favorite images in the #showdown to help make the results even better!
In short:
To generate, for early-member, stable-society and now beta-members: #1100170312106127410 #1103708504142925824 #1100170365604483202 #1100170514670039070 #1080261341362786384 (These channels will auto clear in 30 minute intervals. DM images you'd like saved!)
To vote, for server-wide users: #1100484581037195384
To view top-voted winning generations for server-wide users: #š ļ½pantheon
For chat/bugs/feedback, for early-member, stable-society, and beta-members: #š£ļ½artisan-support-feedback
fluid jungle
ā¼ļø āļø SDXL Bot Updateāļø ā¼ļø
Oh here we are, tried and true <@&1010531326782345307> role.
Are you ready to /dream again?
Weāre now opening our SDXL bot (formerly known to those as "SDXperiment") out to you! For those of you who joined our stage with our AppliedML Team yesterday, you miiiight have caught on that this bot is the latest SDXL checkpoint. *Surprise! *
Growing our pool of users who generate AND vote on these generations is essential to the bettering of our precious little bot through the help of RLHF data and you!
Keep your eyes peeled for when more groups are given wider access! It might be sooner than you think~
Have a great weekend! 
Get started here:
timid smelt
Are you passionate about open source AI?
Are you based in the NYC area?
Come join an IRL meetup hosted by Lightning AI and Stability AI at a secret location in NYC!
š¤ Help democratize AI with collaboration and open source
š Get a chance to show off your AI demos, and support transparent, inclusive, and accessible technology
Find out more at https://partiful.com/e/AqsGTfRFmbCgIlMMnbEk?fbclid=PAAaZWisFhYYilUdLwfY5xhyuDYpaxfG3m6HB7mn4AacS6umhKtxn-8Qs_Wpk
Discord Event: https://discord.gg/stablediffusion?event=1105528321464344657
fluid jungle
ā¼ļø NEW Stable Animation Integration Launch! + 200k Raffle Entryā¼ļø
Lights, Camera, ANIMATION! š¬
@everyone ,
We heard you wanted some new tools in your workshop? š ļø
Come stop by #1098959457964277831 to visit our soft-launch of Stable Animation; where artists and developers alike have the ability to use Stable Diffusion models to generate live-time stunning animations! Create animations purely from prompts, start with an initial image, or drive an animation from a source videoā
*The skyās the limit! *
Get started here:
On a 200k note,
We have now officially opened our #1051259766976040960 channel to celebrate another massive milestone! if you'd like to be entered into our raffle/giveaways running for the entirety of May, then go react to the message in our #1051259766976040960 channel here: #1051259766976040960 message š Best of luck~
More exciting news to look out for soon. . .
fluid jungle
** 200K SDXL-Beta Bot Server-wide LAUNCH!!**
@everyone,
As a part of our 200k celebratory festivities, weāve slowly opened up group-by-group in order to test and tinker with the SDXL Beta Bot. Weāre finally rolling out the red carpet!
The SDXL Beta Bot will now operate SERVER-WIDE in the designated generation channels-- exclusively to this server! With more and more of your RLHF data and preferred aesthetics by voting, weāll become closer and closer to fine-tuning this model into a collaborative enhanced masterpiece!
Please be aware that this bot is in Beta and that this release is intended for research purposes only!
Thanks for being a dazzling part of our community. Have fun!
BOT INSTRUCTIONS
When in the generation channels, you will see the following buttons:
š °ļøVote: To vote for image A, located on the left hand side
š ±ļøVote: To vote for image B, located on the right hand side
āļøDM A and āļøDM B: Clicking these buttons will send the selected image to your DMs, so long as your msg perms arenāt disabled!
šreDo: will regenerate your prompt for you, giving you new outputs
šØreStyle: will regenerate your prompt for you, in a different/randomized style preset that consists of
āAnime, Photographic, Comic Book, Fantasy Artāš¼ļøreSize: will regenerate your prompt for you, giving you a new randomized size dimension of your images
GET STARTED
To generate: Use
/dreamin #1100170312106127410 - #1101178553900478464
These channels auto clear in 30 minute intervals! DM yourself your favorite images!To vote: #1100484581037195384
To view top-voted generations: #š ļ½pantheon
Bot-updates: #1047610792226340935
SDXL Chat: #āØļ½sdxl
ā.ć.:ć»Ā°ā.ć.:ć»Ā°ā.ć.:ć»Ā°ā.ć.:ć»Ā°ā
But wait, there's more! Weāve hit 200K!
Thank you to all of you who have helped celebrate and encourage the rapid paced world of AI and Stable Diffusion. Much love!
As apart of our thanks to you, come join in on some of these upcoming and ongoing festivities:
Ongoing giveaways: Opt in to our giveaways here #1051259766976040960 message
Prompt-Guess!: Our bot master Kevi has created an incredibly addicting game that challenges your prompting skills in a different way than you might be used to! It's an incredible amount of fun and even has a leaderboard you can use to show off your abilities: https://discord.com/channels/1002292111942635562/1103141787960344576
šļø Stable Stage event: Happening this Wednesday at <t:1684346400:t> on the https://discord.com/channels/1002292111942635562/1005545148211527800 Keep your eyes on this even link here https://discord.gg/ZcjG95fm?event=1107712581495439431 for more details
šļø Stable Digest #7: Starring our lovely Animation team, coming out <t:1684429200:F>
š§ **More server/channel updates! ** We've been reviewing your feedback and doing some channel-clean up in the server to keep things a bit more focused and tidy. You may notice a fresh new look as some of our channels have been reorganized and adjusted to better fit what you want!
ā.ć.:ć»Ā°ā.ć.:ć»Ā°ā.ć.:ć»Ā°ā.ć.:ć»Ā°ā
fluid jungle
StableStudio: the Open-Source Future of DreamStudio
<@&1004114999867879506>
Stability AI is thrilled to announce StableStudio, the open-source release of our premiere text-to-image consumer application DreamStudio. StableStudio marks a fresh chapter for our image generation interface and showcases Stability AI's dedication to advancing open-source development within the AI ecosystem! Welcome to this new home for open-source collaboration!
Since the original release of Stable Diffusion, DreamStudio has served as Stability AIās primary interface for its new models and features. Most recently it was upgraded to use SDXL Beta, our latest image generation model. The application was first conceived as a platform to bring generative AI capabilities to the masses. Since then, millions of images have been created by its users.
We believe the best way to expand upon that impressive reach is through open, community-driven development, rather than private iteration on a closed-source product.
We're eager to see what the community will create! If youāre interested in getting involved, check out the links below!
GET STARTED:
StableStudio channel: #1108055793674227782
Github:https://github.com/Stability-AI/StableStudio
README:https://github.com/Stability-AI/StableStudio/blob/main/packages/stablestudio-ui/README.md
Blogpost: https://stability.ai/blog/stablestudio-open-source-community-driven-future-dreamstudio-release
Stable-Stage Event TODAY: https://discord.com/events/1002292111942635562/1107712581495439431
fluid jungle
āļø Clipdrop launches Reimagine XL! āļø @everyone
Clipdrop proudly presents an upgraded version of Reimagine ā the ultimate image reimagination tool! š ļø
This upgraded version, exclusively available on Clipdrop, takes your image reimagining capabilities to the next level by removing direct pixel sourcing from the original image, ensuring bold and unique output with every use.
Reimagine XL takes what Stable Diffusion Reimagine started and goes even further. We've trained this model using larger image resolutions, resulting in higher quality results. It particularly excels at handling various aspect ratios, making panoramic landscapes breathtakingly grand and close-up shots incredibly detailed.
The best part?
You don't need any technical know-how to use it!
Simply upload an image and let Reimagine XL work its magic. Its advanced image encoder technology ensures that you get a unique and diverse output every time.
Whether you're redecorating your bedroom, creating fashion sketches, or changing a landscape, Reimagine XL has got you covered. See for yourself!
INFO & GETTING STARTED:
About Stable Diffusion Reimagine: https://stability.ai/blog/stable-diffusion-reimagine
Clipdropās Reimagine XL: https://clipdrop.co/stable-diffusion-reimagine
NEW Blogpost on Reimagine XL: https://stability.ai/blog/stability-ai-clipdrop-launches-reimagine-xl
Youtube Video: https://www.youtube.com/watch?v=2dgBf-hEPDQ
fluid jungle
š **Clipdrop launches Uncrop: The Ultimate Aspect-Ratio Editor**š
@everyone
Clipdrop is excited to announce Uncrop,
its AI-generated āoutpaintingā tool that allows users to change the ratio of an image by creating an expanded background to complement ANY existing photo or image! š¼ļø
Built on Stability AI's text-to-image model Stable Diffusion XL, Uncrop allows users to adjust dimensions by reconstructing images and expanding their visual canvas! šØ
With just a few simple steps, users can upload a cropped or imperfect image and witness the magic of outpainting technology in action. Using advanced algorithms and deep learning techniques, it analyzes the content within the uploaded image and generates a visually plausible representation of what could have been there. š
Uncrop can assist users in multiple cases! To change the composition of a photo after having taken it, Uncrop can help adjust the framing to make it look more appealing. To adapt an imageās ratio to display on a banner or a large canvas, Uncrop will resize and adjust the image to fit the dimensions without distorting or losing important details. š
Try it for free in the Clipdrop web app--with no need to log in!
Get Started Here!
ā Clipdrop Uncrop: https://clipdrop.co/uncrop
ā Clipdrop Video: https://www.youtube.com/watch?v=aHIHx1ZZ9OQ
ā Blogpost: https://stability.ai/blog/clipdrop-launches-uncrop-the-ultimate-aspect-ratio-editor
fluid jungle
Update on SDXL Open Release: SDXL 0.9
Who doesn't love a little countdown? ā°
We're thrilled to announce that ā¼ļøSDXL 0.9 is HEREā¼ļø
We couldn't contain our excitement and decided to give an exclusive early access preview to select academic researchers, before rolling out the general release red carpet to YOU; our community!
We are steps away from SDXL being open for all! With the help of these select researchers feedback, they will help us make our public 1.0 release the best it can be for you!
As SDXL is visiting our researchers, we invite all our users to take a peek behind the curtains of SDXL 0.9 on ClipDrop! You can also tinker away on the previous research preview on the server in #1100170312106127410 - #1101178553900478464. Dive in if you haven't already, experiment, and get a glimpse of what capabilities SDXL has to offer while awaiting the 1.0 full release in the coming weeks!
(Please be aware no changes have been made to the bot currently on the server!)
The team has been working endlessly to get to this point, and your feedback, patience, and kindness is and has been immensely valued! 
Join us at our Stable-Stage with the one and only @surreal pollen to talk about this exciting new installment, ask all your ever-burning curious questions, and tune in to what steps lie ahead! š š
.ć . ā¢ ā . Ā° .ā¢ Ā°:. *ā Ā° . ā
Let's prepare for takeoff together!
INFO
timid smelt
Hello, @everyone!
My apologies for the ping ā¤ļø
If you have not already, I ask you please take a moment to fill out our current Stable Foundation Community Feedback Survey (est. 1-2 minutes to complete).
These surveys and the results we collect from them are incredibly helpful when planning the future of the server itself. By using your responses, we can more closely monitor how our recent changes have been received as well as confirming our direction matches your expectations. We may not be able to give everyone exactly what they want, but we can use surveys like these to help get as close as we possibly can!
What to expect:
Similar to the last survey, the first section will focus on all the different activities and events we host for the community to participate in. Having your feedback about these events help us tune them into something that you'll find both interesting and also rewarding of the time you invest into them!
The second half of the survey focuses on how the server feels and looks to you while using it. We use this to help drive further changes to our channel structure and lineup to make sure you can always find what you're looking for as easily as possible!
(This is why it's important for even new members of the community to give us their honest thoughts!)
You are welcome to leave your feedback anonymously if you choose to do so!
Simply leave the first section "Discord Username (Optional)" blank and then hit the "Next" button at the bottom left!
Please be as detailed as you can and don't hold back on any of the excellent feedback or ideas we know you have!
Finally, a huge Thank You to everyone who takes a few moments out of their day to make this place even better!
https://docs.google.com/forms/d/1J96YgbX69fP9h3QXYaSFx1jYmJVkBBp47VQ_ucjZusU
fluid jungle
āļø__Bot Update: SDXL 1.0 Release Candidates!!__āļø
@everyone, Uhoh.. it's the big bad /everyone/ ping..
Ready to kick off the week?
Weāre hitting our SDXL Bot on our server with some very new and exciting updates! At long last, we finally have all of our final release candidates for SDXL 1.0ā now āØLIVEāØ on the bot!
Just as last time, the bot will give you randomized sets of parameters, samplers, aspects, and outputs!
Come hop on over to our bot channels, generate some amazing images, and vote on your favorites, so that YOU; the community can help us crown our #1 release candidate for the big SDXL 1.0 release! š
No big deal or anything, right?
šø š© š½ š± ā ā ā 
āļø GET STARTED āļø
fluid jungle
Clipdrop Launches: Stable Doodle!
@everyone, Meet
Stable Doodle,a sketch-to-image tool that converts a simple drawing into a dynamic image, providing limitless imaging possibilities to professionals and hobbyists alike! āļø
Enhance a range of industries, from education, to creative design, to fashion and the arts. Stable Doodle is available to try for FREE on the Clipdrop by Stability AI website along with SDXL 0.9.
This user-friendly approach enables designers, illustrators, and other professionals to free up valuable time and maximize efficiency. Ideas drawn as sketches can be immediately implemented into works to create designs for clients, material for presentation decks and websites, or even create logos. The possibilities are endless! Give it a whirl!
GET STARTED! ā“
Stable Doodle: https://clipdrop.co/stable-doodle
fluid jungle
Hello, @everyone! 
As we approach the finish line, we have another exciting update for you regarding SDXL and the bot!
Thanks to all of your amazing voting efforts, the bot has generated over 2,746,114 images with over 941,313 nominees to the Showdown to date (wow)!
You have genuinely blown us away by your incredible participation & we can now announce our top three finalists have been selected!
In order to crown our ultimate champion, the finalists have been loaded into StableDreamer and will be randomly used when generating images. š
We have also added in something our keen testers have been asking for: Second stage!
What does "second stage" mean? It means after this update things will look even better!
We continue to listen to all of your feedback and want to take a moment to say how much we appreciate all of your patience and excitement surrounding the launch of SDXL.
We really hope you'll enjoy testing out the latest update in preparation for the open source release of SDXL
Generate, Play, and Vote Now!
#āØļ½sdxl / word from Emad: #āØļ½sdxl message
#1100170312106127410 - #1101178553900478464
#š£ļ½artisan-support-feedback
lyric bobcat
fluid jungle
SDXL 1.0 Release & Super Stage!!
@everyone
Stability is proud to announce the release of SDXL 1.0; the highly-anticipated model in its image-generation series! After you all have been tinkering away with randomized sets of models on our Discord bot, since early May, weāve finally reached our winning crowned-candidate together for the release of SDXL 1.0, will now be available via Github, DreamStudio, API, Clipdrop, and AmazonSagemaker!
Your help, votes, and feedback along the way has been instrumental in spinning this into something truly amazingā It has been a testament to how truly wonderful and helpful this community is!
For that, we thank you! ā¤ļø
SDXL has been tested and benchmarked by Stability against a variety of image generation models that are proprietary or are variants of the previous generation of Stable Diffusion. Across various categories and challenges, SDXL comes out on top as the best image generation model to date.
Some of the most exciting features of SDXL include:
Freedom of expression: Best-in-class photorealism, as well as an ability to generate high quality art in virtually any art style. Distinct images are made without having any particular āfeelā that is imparted by the model, ensuring absolute freedom of style
š” Enhanced intelligence: Best-in-class ability to generate concepts that are notoriously difficult for image models to render, such as hands and text, or spatially arranged objects and persons (e.g., a red box on top of a blue box)
Simpler prompting: Unlike other generative image models, SDXL requires only a few words to create complex, detailed, and aesthetically pleasing images. No more need for paragraphs of qualifiers.ā More accurate: Prompting in SDXL is not only simple, but more true to the intention of prompts. SDXLās improved CLIP model understands text so effectively that concepts like āThe Red Squareā are understood to be different from āa red squareā. This accuracy allows much more to be done to get the perfect image directly from text, even before using the more advanced features or fine-tuning that Stable Diffusion is famous for.
š All of the flexibility of Stable Diffusion: SDXL is primed for complex image design workflows that include generation for text or base image, inpainting (with masks), outpainting, and more. SDXL can also be fine-tuned for concepts and used with controlnets. Some of these features will be forthcoming releases from Stability.
Come join us on stage with Emad and Applied-Team <t:1690394400:R> for all your burning questions! Get all the details LIVE!
GET STARTED (Links open <t:1690399800:R>)
fluid jungle
Introducing: Stable Code! 
@everyone
Stability AI has just announced the release of StableCode, its
This product is designed to assist programmers with their daily work while also providing a **great learning tool for new developers **ready to take their skills to the next level!
āļø StableCode offers a unique way for developers to become more efficient by using **three different models ** to help in their coding.
āļø The base model was first trained on a diverse set of programming languages from the stack-dataset (v1.2) from BigCode and then trained further with popular languages like Python, Go, Java, Javascript, C, markdown and C++.
āļø In total, we trained our models on 560B tokens of code on our HPC cluster.
āļø After the base model had been established, the instruction model was then tuned for specific use cases to help solve complex programming tasks.
~120,000 code instruction/response pairs in Alpaca format were trained on the base model to achieve this result.
Check it out yourself!
GET STARTED
Base Model: https://huggingface.co/stabilityai/stablecode-completion-alpha-3b-4k
Instruction Model: https://huggingface.co/stabilityai/stablecode-instruct-alpha-3b
Read more about Stable Code on the Blogpost: https://stability.ai/blog/stablecode-llm-generative-ai-coding
#1098025024541167646
fluid jungle
Japanese StableLM, Marking Entry into International Language Model Market!šŖ
Today, Stability AI released its first Japanese language model (LM), Japanese StableLM Alpha, the best-performing openly available LM created for Japanese speakers.
āļø Japanese StableLM is a 7 billion-parameter general-purpose language model. It stands as the top-performing publicly available Japanese language model, according to a benchmark suite against four sets of other Japanese LMs.
āļø Japanese StableLM Base Alpha 7B will be released under the commercially available Apache License 2.0.
āļø Japanese StableLM Instruct Alpha 7B is a model created for research purposes and is released exclusively for research use.
š¤ For more details, please refer to the Hugging Face Hub page!
MORE INFO
Read more on the Blogpost: https://stability.ai/blog/stability-ai-new-jplm-japanese-language-model-stablelm
Base Model: https://huggingface.co/stabilityai/japanese-stablelm-base-alpha-7b
Instruction Model: https://huggingface.co/stabilityai/japanese-stablelm-instruct-alpha-7b
fluid jungle
Stable Chat: research preview and participation in DEF CON AI Village!
On
July 21st, 2023, we released a powerful new open access large language model. At the time of launch, it was the best open LLM in the industry, comprising intricate reasoning and linguistic subtleties, capable of solving complex mathematics problems and similar high-value problem-solving.
āļø Welcome in the launching of the Stable Chat research preview, a new web interface that empowers the AI community to evaluate our large language models interactively.
āļø Through Stable Chat, researchers can provide feedback on the safety and quality of its responses and flag biased or harmful content to help us improve these open models.
As part of our efforts at Stability AI to build the world's most trusted language models, we've set up a research-purpose-only website to test and improve our technology. We will continue to update new models as our research progresses rapidly. We ask that you please avoid using this site for real-world applications or commercial uses.
Come try it out! Users can create a free account or log in using a Gmail account. 
MORE INFO
https://research.stability.ai/chat
Read more: https://stability.ai/blog/stable-chat-research-defcon-ai-village
fluid jungle
Meet Stable Beluga 1 and Stable Beluga 2, Our Large and Mighty Instruction Fine-Tuned Language Models! š³
Stability AI and its CarperAI lab proudly announce Stable Beluga 1 and its successor Stable Beluga 2 (formerly codenamed FreeWilly), two powerful new, open access,** Large Language Models (LLMs)**. Both models demonstrate exceptional reasoning ability across varied benchmarks!
Stable Beluga 1 leverages the original LLaMA 65B foundation model and was carefully fine-tuned with a new synthetically-generated dataset using Supervised Fine-Tune (SFT) in standard Alpaca format. 
Similarly, Stable Beluga 2 leverages the LLaMA 2 70B foundation model to achieve industry-leading performance.
Both models are research experiments and are released to foster open research under a non-commercial license.
š«§ __MORE INFO:__š«§
#1134212604173242428
Read More: https://stability.ai/blog/stable-beluga-large-instruction-fine-tuned-models
Stable Beluga 1: https://huggingface.co/stabilityai/StableBeluga1-Delta
Stable Beluga 2: https://huggingface.co/stabilityai/StableBeluga2
CarperAI: https://carper.ai/
timid smelt
Hello, <@&1004116145319710840>(s)! š
We have an exciting announcement to send you off into your weekends with:
The amazing team at 
CarperAI and StabilityAI have just released DRLX! š„
DRLX is a new tool to improve diffusion image models (Like Stable Diffusion) using reinforcement learning!
It can improve the generated images' aesthetic scoring, compressibility, and more, so I highly suggest checking it out to see it's capabilities for yourself!
The best part? You can get it right now!:
https://github.com/CarperAI/DRLXRead the docs here: https://drlx.readthedocs.io/
PyPI: https://pypi.org/project/drlx/0.0.2/
fluid jungle
Stable Diffusion's 1 Year Anniversary of Innovation & Community!
Dearest, @everyone
We've reached the beautiful milestone of our
Year Anniversary!
This technology has come an incredibly long way, hasn't it? So has our ever-growing and talented community!
You are the driving force behind this remarkable time in tech and, well.....history! Celebrate with us for the entirety of this month ahead! Dates will be announced as ready!
Exciting Events Ahead:
š Stable-Stage: StableChat & LLM Research:
We had our kick-off stage yesterday, filled with insightful discussions about StableChat and the latest developments in Language and Learning Model research! No worries if you missed it! We'll be uploading to our Youtube (https://www.youtube.com/@stablefoundationai), and have more stages underway for this month!
š Month-long Prizes & Giveaways:
Get ready for a celebration of giving! Throughout this special month, we'll be giving out prizes for our valued community members.
āļø CogX x Stability.AI:
Dive into the fascinating intersection of AI and innovation as we participate in the renowned CogX event! More here: https://discord.gg/wT3pz9ZJ?event=1143239071443988481
š“ Stable-Stage: Invoke:
An engaging and interactive session awaits to evoke your imagination! Stay tuned for more information and keep your eyes peeled in our Events section at the top of the server!
š°ļø Stable Diffusion: Through the Ages Event!
Step into history as we explore the evolution of technology and its impact on our lives. Through the Ages invites you to craft a series of three images using YOUR favorite Stable Diffusion models from the past year. You have the flexibility to utilize any custom model from the community or leverage a strong foundational model like SDXL.
š® Sticker-Scrapbooking Event:
Unleash your creativity in the Sticker-Scrapbooking Event, where your designs could become cherished symbols of our vibrant community...and come together to create a sticker book?!
š”Stable-Society Spotlights:
Discover the incredible stories and achievements of our community members in the Stable-Society Spotlights...More to come!
timid smelt
Hey all, sorry for the influx of sudden pings. I'm actively trying to figure out why they keep duplicating as we are not initiating them from our side.
We have escalated a ticket to Discord for assistance
fluid jungle
Introducing Stable Audio: Stability.AIās first AI product for music and sound generation! š§
<@&1004115656620392538> & <@&1004116145319710840>
āReady for something new? Welcome in Stable Audio!ā
Stable Audio is a first-of-its-kind product that uses the latest generative AI techniques to deliver faster, higher-quality music via an easy-to-use web interface. Stability AI offers both a basic free version of Stable Audio, which can be used to generate and download tracks of up to 20 seconds, as well as āProā subscription, which delivers 90-second tracks that are downloadable for commercial projects.
Stable Audio is the first music generation product enabling the creation of high-quality, 44.1 kHz music for commercial use via latent diffusion. The latent diffusion architecture uses audio conditioned on text metadata as well as audio file duration and start time, allowing for control over the content and length of the generated audio.
We welcome you to give Stable Audio a whirl for yourself, and provide us your feedback! We want to hear from YOU!
GET STARTED:
fluid jungle
Stable Audio: Prompt to Audio
Have you tried out our newest prompt to audio product; Stable Audio, @everyone ?
Enter a world in which you can create full on music, samples, and SFX all in a matter of seconds! Streamline your workflows, create new ones, and explore the infinite possibilities of audio in the AI realm!
Try out these prompts showcased in our official launch video!
-
Trance, Ibiza, Beach, Sun, 4 AM, Progressive, Synthesizer, 909, Dramatic Chords, Choir, Euphoric, Nostalgic, Dynamic, Flowingšµ -
Ambient house, new age, meditation, advertisement, 808 drum machine, 808 kick, claps, shaker, synthesizer, synth bass, lead heavily reverbed, beautiful, inspiring, modern, sleek, beautiful, inspiringšµ -
Post-Rock, Guitars, Drum Kit, Bass, Strings, Euphoric, Up-Lifting, Moody, Flowing, Raw, Epic, Sentimental, 125 BPMšµ
GET STARTED:
fluid jungle
Typed.Visions SDXL Promptbook āØļø
<@&1072165187634860163> & <@&1004114999867879506>
We're thrilled to share some fantastic work with you all today! Introducing our sparkly new and official SDXL Promptbook, made by our prompt-smith @flint osprey -- a valuable resource designed for all, especially catering to beginners, or who simply need a refresh on all of the AI lingo. Beginner-friendly vocabulary and definitions, visual explanations, and beautiful example prompts to try-- ready to give it a peek?
š : https://docs.google.com/presentation/d/1HEcE3qOAGVujcDaNQbiLXyx7zwKHQkXEILsYBhsot7A/edit?usp=sharing
: 
fluid jungle
EU AI Act Update! š° @everyone
Next week, the EU will enter a final sprint to decide the fate of AI models under the AIAct. We've been meeting with EU leaders all through the summer, fighting to safeguard the everyday developers & researchers who share and contribute to open models.
š§ Issue at Hand--
The latest draft of the Act will treat all models identically once they are released publicly. It doesn't matter if they are powerful base models from corporate labs or fine-tuned models shared by DIY tinkerers. These "one size fits all" rules were designed for sophisticated corporate actors, and they will have a serious chilling effect on everyday developers, independent researchers, and small businesses across Europe who share and contribute to open models..
š¤ Problematic Impact?
Grassroots innovation in open models enhances AI safety and accessibility. The proposed rules risk stifling this by hindering researchers and developers, jeopardizing the EU's competitive edge in AI.
ā© Way Forward..
To lead in AI, the EU must nurture an ecosystem that supports grassroots experimentation. Some Member States propose stripping back model provisions entirely. Failing that, we urge a careful distinction between open models based on their release intent and deployment.
š Stay Informed!
Check out our detailed comments on the latest draft here: https://static1.squarespace.com/static/6213c340453c3f502425776e/t/65140501f63cd217acb2d780/1695810818768/Grassroots+innovation+in+open+models+ā+suggested+amendments+to+the+AI+Act.pdf
Not sure what the AI Act is? Read even more here: https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence
DreamStudio & API Survey āļø
Thank you for being an integral part of our vibrant community. Your insights play a crucial role in shaping the future of Stability AI, and we're excited to invite you to share your thoughts with us!
Help us improve our products by leaving valuable feedback and your insight: https://docs.google.com/forms/d/e/1FAIpQLSdZsNWDDYEgLXybGNj_a_hIGBHYTBJB9DzuC9CnceB4u8z8zw/viewform
Thank you so much!
fluid jungle
Introducing Stable LM 3B: Bringing Sustainable, High-Performance Language Models to Smart Devices!!
@everyone
Today, we proudly launch an experimental version of
Stable LM 3B, the latest in our suite of high-performance generative AI solutions. At 3 billion parameters, Stable LM 3B is a compact language model designed to operate on portable digital devices like handhelds and laptops, and weāre excited about its capabilities and portability!
Compact and Portable Design:
Stable LM 3B, at 3 billion parameters, is designed for portable digital devices such as handhelds and laptops.
Compact size makes it highly portable and versatile for a variety of applications.
Efficiency and Affordability:
Requires fewer resources and boasts lower operating costs compared to larger models.
Affordability and efficiency make it accessible to a broad user base.
Environmentally friendly, as it consumes significantly less power than larger models.
High Competitive Performance:
Outperforms previous state-of-the-art 3B parameter language models.
Competitive with some of the best open-source language models at the 7B parameter scale.
Broadened Application Range:
Enables the development of cutting-edge technologies with strong conversational capabilities on edge devices and home PCs.
Ideal for applications like creative writing assistance with low costs and high performance.
Improved Downstream Performance:
Significantly better at producing text while maintaining fast execution speed.
Enhanced downstream performance on natural language processing benchmarks, including common sense reasoning and general knowledge tests.
GET STARTED:
Read More on the Blogpost: https://stability.ai/blog/stable-lm-3b-sustainable-high-performance-language-models-smart-devices
Github: https://github.com/Stability-AI/StableLM#stablelm-alpha-v2
HF: https://huggingface.co/stabilityai/stablelm-3b-4e1t
Want to provide info/feedback?: research@stability.ai
fluid jungle
Happy Friday~!
<@&1004114999867879506>
Today we are excited to present a new collaboration to help launch your ideas with Stability AI!
Our partners at Backdrop Build are offering an exciting new opportunity for builders in crypto and AI: A free, async friendly, 4 week program with over $50k in grants for top projects!
The program will run from Nov 9th to Dec 3rd and is open to all with minimal time commitment!
Are your ideas still in the planning phase? No problem! Backdrop will provide the guidance you need to bring your project to the next level and expand your own personal skills!
Apply here before Oct 29: https://backdropbuild.com/apply
fluid jungle
Stability AI Previews Enhanced Image Offerings For Businesses:
Improved Quality, APIs & Product Features 
Stability AI is excited to share updates regarding our next generation text-to-image products, featuring private previews and a public beta of upcoming offerings, including enterprise-grade image services, APIs and enhanced image capabilities.Today, we are rolling out the red carpet for:
šŖ Sky Replacer:
https://clipdrop.co/real-estate/sky-replacer
Sky Replacer, now offered on Clipdrop, is an exciting new tool that allows users to replace the color and aesthetic of the sky in their original photos with a selection of nine alternatives to improve the overall look and feel of the image.
Stable 3D Private Preview:
Meet Stable 3D, an automatic process to generate concept-quality textured 3D objects that eliminates much of that complexity and allows a non-expert to generate a draft-quality 3D model in minutes, by selecting an image or illustration, or writing a text prompt.
š ļø Stable FineTuning Private Preview
The long-awaited Stable FineTuning provides enterprises and developers the ability to fine-tune pictures, objects, and styles at record speed, all with the ease of a turnkey integration for their applications.
š Read more: https://stability.ai/blog/stability-ai-enhanced-image-apis-for-business-features
timid smelt
**Donāt miss out! **
Today is the last day for an incredible opportunity with Backdrop Build: Offering a 4 week program to support your ideas to launch something on Stability!
Free, Async friendly, and with over 50k in grants for top projects, thereās no reason to wait! Sign up at http://backdropbuild.com
#š£ļ½announcements message
fluid jungle
Stable Video Diffusion! šļø
@everyone!
Today, we have released (in research preview) our newest initiative in Generative AI technology: Stable Video Diffusion!
āÆļø Researchers are invited to experiment with Stable Video Diffusion with the code for Stable Video Diffusion now available on our GitHub repository. The weights required to run the model locally can be found on our HuggingFace page (Note: 40GB of VRAM required) & further details regarding the technical capabilities of the model can be found in our research paper. Additionally, you can join the waitlist for the upcoming web experience including a text-to-image interface coming soon.
š¬ As part of this launch, weāve opened up a #ā¶ļ½stable-video-diffusion channel here on the server & encourage you all to share your creations as well as your incredibly valuable feedback while you test out what itās capable of!
While we eagerly update our models with the latest advancements and work to incorporate your feedback, we emphasize that this model is not intended for real-world or commercial applications at this stage. Your insights and feedback on safety and quality are pivotal to refining this model for its eventual commercial release.
MORE:
#ā¶ļ½stable-video-diffusion
Blogpost: https://stability.ai/news/stable-video-diffusion-open-ai-video-model
Join the Waitlist:https://stability.ai/contact
Research Paper: https://stability.ai/research/stable-video-diffusion-scaling-latent-video-diffusion-models-to-large-datasets
Github: https://github.com/Stability-AI/generative-models
fluid jungle
Introducing SDXL Turbo! 
@everyone
š Built on the same technological foundation as SDXL 1.0, SDXL Turbo features the enhancements of a new technology: Adversarial Diffusion Distillation (ADD).
šØ SDXL Turbo accelerates image generation,* delivering high-quality outputs* within notably shorter time frames by decreasing the standard suggested step count from 30, to 1! To try out the model right away, visit Stability AIās image editing platform Clipdrop, demonstrating the real-time text-to-image generation capabilities!
ā¼ļø *This model is being released under a non-commercial research license and permits personal, non-commercial use. *
šØļø If you are interested in using this model for commercial purposes, please contact us at https://stability.ai/contact to learn about our self-hosted commercial license options.
š° Stay updated on our progress by signing up for our newsletter! (https://stability.ai/home#newsletter)
INFO:
#1178805966335119472
Model weights/code on Hugging Face: https://huggingface.co/stabilityai/sdxl-turbo
Clipdrop SDXL Turbo: https://clipdrop.co/stable-diffusion-turbo
Blogpost: https://stability.ai/news/stability-ai-sdxl-turbo
fluid jungle
Hack{hers} Hackathon in Silicon Valley! Dec. 3rd
Buckle up for a community creative event accompanied by some of the top female builders in Silicon Valley!
šļø Be joined by our lovely lineup of speakers and judges-- Kelsey Szot, Co-founder of Adept, Keerthana Gopalakrishnan, Research Engineer at DeepMind, Christina Kim, Member of Technical Staff at Openai, and Puja Rajan, Member of Technical Staff at Anthropic!
ā¼ļø Apply now and have the opportunity to share your skills this Sunday in the Bay Area!
fluid jungle
Create and monetize your AI art style | LinkedIn
<@&1072165187634860163>
Want to learn how to monetize your creations in this space? Join us on LinkedIn for an event hosted by Artificial Inspiration <t:1701968400:F> with our very own @vestal island , as he speaks on the best outlets to empower your art!
Whether you're a photographer, illustrator, AI artist, or even a casual user-- this is your ticket to not only craft, but also monetize your unique style.
Mark your calendars, and let's innovate!
š https://www.linkedin.com/events/7134312037714677761/about/
šļø https://discord.gg/stablediffusion?event=1181717838952661052
fluid jungle
Introducing StableLM Zephyr 3B: A New Addition to StableLM, Bringing Powerful LLM Assistants to Edge Devices š£
@everyone
Today we are happy to announce the release of StableLM Zephyr 3B: a new 3 billion parameter chat model! This model represents the latest in our series of lightweight LLMs, preference tuned for instruction following and Q&A type tasks.
60% smaller than 7b models, Stable Zephyr allows accurate, and responsive output on a large variety of devices without requiring high end hardware!
š This model is released under a non-commercial research license that permits personal, non-commercial use.
Read our blogpost below for more info!
MORE INFO:
Model code & weights: https://huggingface.co/stabilityai/stablelm-zephyr-3b
Read more on the Blogpost: https://stability.ai/news/stablelm-zephyr-3b-stability-llm
Discussion: #1098025024541167646
fluid jungle
Stability releases: Stable Zero 123 !
@everyone
Today we are releasing Stable Zero 123, a new in-house generative AI model from Stability AI capable of creating 3D objects from any input image!
š ļø Based on Zero123, this model was trained using a filtered database from Objaverse with notably improved results compared to previous 3D models
Note: This model uses similar amounts of resources to Stable Diffusion 1.5 to generate 1 novel view. For the generation of 3D objects, a minimum of 24GB VRAM is recommended.
ā¼ļø Stable Zero123 is being released for non-commercial and research use.
MORE INFO
Discussion around release: #1184555609232777337
Read more here: https://stability.ai/news/stable-zero123-3d-generation
Weights: https://huggingface.co/stabilityai/stable-zero123
timid smelt
@everyone
Today we are officially introducing the Stability AI Membership!
Our goal with this membership is to sustainably provide state-of-the-art technologies that promote innovative and accessible technological advancements for all.
First and foremost, the base level of the Stability AI Membership will remain free and open to all non-commercial uses for individuals, researchers, and creatives. We continue to strongly believe that our core model releases should always be provided to the community for free, avoiding prohibiting barriers of entry into AI technology and enabling competition and research efforts to drive technology forward as it always has.
For commercial use of our core models, there will be two options:
For creators, developers, startups, earning less than $1m in annual revenue, less than $1m in institutional funding, and with less than 1m active users, the Professional tier of the Stability AI Membership will be implemented at $20 a month. We additionally understand the importance of taking into account the impact that geographical and personal situations may have on the affordability of this tier. We encourage anyone who finds this cost a prohibitive factor to contact us with your proposal so we can review what options are available to assist you with available discretionary grants.
Finally, for larger entities, our Enterprise tier allows for custom pricing and additional options for enterprise support and solutions engineering to help you deploy and manage our models.
As our company continues to mature and scale, the Stability AI Membership will play a pivotal role in funding our future research and development of core models. This enables us to lay the foundation for builders, creatives, and companies at the forefront of driving transformational change, powered by Stability AI.
For more information about the Stability AI Membership and our commitment to open and accessible AI models, read our full announcement here: https://bit.ly/3t792TG
Visit our membership page here: https://stability.ai/membership
fluid jungle
Emad AMA! šļø
@everyone
Come join us with Emad for an AMA!
He will be answering your questions about the Stability AI Membership, our models and products, or anything Gen AI!
Listen in and ask questions. We canāt wait to catch up!
fluid jungle
Stable Diffusion 3 Waitlist
š° Announcing Stable Diffusion 3, our most capable text-to-image model, utilizing a diffusion transformer architecture for greatly improved performance in multi-subject prompts, image quality, and spelling abilities.
āļø Today, we are opening the waitlist for early preview. This phase is crucial for gathering insights to improve its performance and safety ahead of open release.
You can sign up to join the waitlist and learn more here: https://bit.ly/3UQ0HPE
fluid jungle
Stable Diffusion 3 Research Paper š°
Today, we're publishing our research paper that dives into the underlying technology powering Stable Diffusion 3, outperforming state-of-the-art text-to-image generation systems such as DALLĀ·E 3, Midjourney v6, and Ideogram v1.
Learn more here: https://bit.ly/4bULWkF
fluid jungle
Stable Video 3D š„½
Today, we are releasing Stable Video 3D, a generative model based on Stable Video Diffusion. This new model advances the field of 3D technology, delivering greatly improved quality and multi-view.
Learn more here: https://stability.ai/news/introducing-stable-video-3d
fluid jungle
Stable Diffusion 3 Live on API š
Today, we are pleased to announce the availability of Stable Diffusion 3 and Stable Diffusion 3 Turbo on the Stability AI Developer Platform API.
We have partnered with Fireworks AI, the fastest and most reliable API platform in the market, to deliver these models.
In keeping with our commitment to open generative AI, we aim to make the model weights available for self-hosting with a Stability AI Membership in the near future.
You can get started and learn more here: https://bit.ly/3xHrtjG
Prompt: Awesome artwork of a wizard on the top of a mountain, he's creating the big text "Stable Diffusion 3 API" with magic, magic text, at dawn, sunrise.
timid smelt
@everyone
Introducing Stable Artisan
Ready for the return of the bot? Today, weāre introducing Stable Artisan; a multimodal generative AI Discord bot that provides users the ability to use Stability AI's API offerings within the Discord ecosystem.
Being fully multimodal, Stable Artisan will provide image generation via Stable Image Core and Stable Diffusion 3, and video generation via Stable Video Diffusion.
You can expect more updates in the coming week as we continue to introduce more exciting features!
Read the full announcement here: https://bit.ly/4aiVy6C
Get started:
#artisan-faq
#1237459938901491852 #artisan-1 #artisan-2 #artisan-3 #1237460438229450772
fluid jungle
Stable Assistant Updates
@everyone
Weāre excited to announce new editing features in Stable Assistant, our user-friendly chatbot. Leveraging the power of Stable Diffusion 3, these advanced text-to-image capabilities produce higher quality outputs.
Try it for free on your images: https://stability.ai/stable-assistant
fluid jungle
Stability AI x HUG
Join us with HUG at the Innovation Laboratory, a 4-week guided summer course on how to train your own AI model, showcasing your creative abilities in multimodal generative AI!
You can register in advance for this event here: https://www.studios.thehug.xyz/lab
fluid jungle
@everyone
The āweightā is nearly over! Today, at Computex Taipei, our Co-CEO, Christian Laforte, officially announced the open release date of Stable Diffusion 3 Medium for June 12th.
š Sign up to the waitlist to be the first to know when the model releases: https://stability.ai/stablediffusion3
fluid jungle
Watch the full announcement on the Stable Diffusion 3 Medium weight release at Computex Taipei: https://youtube.com/watch?v=MCi8jgALPYA š
fluid jungle
Stable Audio Open is here!
<@&1004115656620392538>
Weāre excited to announce Stable Audio Open, an open source model optimised for generating short audio samples, sound effects and production elements using text prompts. This release marks a key milestone as we further open portions of our generative audio capabilities to empower sound designers, musicians and creative communities.
Trained on data from FreeSound and Free Music Archive, Stable Audio Open provides free model weights to download on HuggingFace, documentation & code examples, and local fine-tuning on your custom audio data.
Key features include:
šµGenerating up to 47 seconds of high-quality audio clips.
š„Creating drum beats, instrument riffs, foley and ambient sounds.
š§Enabling variations and style transfer of uploaded samples.
Learn more and get started here: https://stability.ai/news/introducing-stable-audio-open
fluid jungle
Stable Diffusion 3 Medium Open Weights @everyone
Today, weāre thrilled to announce the open weights for Stable Diffusion 3 Medium, the latest and most advanced text-to-image AI model in our Stable Diffusion 3 series! š
This new release represents a major milestone in the evolution of generative AI and continues our commitment to democratising this powerful technology.
SD3 Medium is a 2 billion parameter SD3 model that offers some notable features:
š· Overall Quality and Photorealism: Delivers images with exceptional detail, color, and lighting, enabling photorealistic outputs as well as high-quality outputs in flexible styles. Success in addressing common pitfalls of other models, such as realism in hands and faces, is achieved through innovations such as the 16-channel VAE.
š¼ļø Prompt Understanding: Comprehends long and complex prompts involving spatial reasoning, compositional elements, actions, and styles. By utilizing all three text encoders or a combination, users can trade off performance for efficiency.
āļøTypography: Achieves unprecedented text quality with fewer errors in spelling, kerning, letter forming, and spacing by leveraging our Diffusion Transformer architecture.
š„ļøResource-efficient: Ideal for running on standard consumer GPUs without performance degradation, thanks to its low VRAM footprint.
š§² Fine-Tuning: Capable of absorbing nuanced details from small datasets, making it perfect for customisation.
š Learn more and get started here: http://stability.ai/news/stable-diffusion-3-medium
fluid jungle
License Update @everyone
At Stability AI, weāre committed to releasing high-quality Generative AI models and sharing them generously with our community of innovators and media creators.
We acknowledge that our latest release, Stable Diffusion 3 Medium, didnāt meet our communityās high expectations, and recognize that the commercial license originally associated with SD3 caused some confusion and concern.
Weāve made a few improvements to address these concerns.
Read the full update here: http://stability.ai/news/license-update
fluid jungle
Stable Video 4D
We are pleased to announce the availability of Stable Video 4D, our very first video-to-video generation model that allows users to upload a single video and receive dynamic novel-view videos of eight new angles, delivering a new level of versatility and creativity.
In conjunction with this announcement, we are releasing a comprehensive technical report detailing the methodologies, challenges, and breakthroughs achieved during the development of this model.
Learn more about this release and access the report here: https://stability.ai/news/stable-video-4d
fathom fossil
Hi,
We have added a new command, /style, to Stable Artisan. By executing this command and specifying a style image and prompt, you can generate an image that matches the specified style. You can create images like a Van Gogh-style cat or a Japanese-style spaceship.
Pasted an example drawing of a Van Gogh-style cute cat.
Please enjoy!
fluid jungle
Stable Fast 3D Launch š
<@&1072166265982034011>
We are excited to introduce Stable Fast 3D, Stability AIās latest breakthrough in 3D asset generation technology. This innovative model transforms a single input image into a detailed 3D asset in just 0.5 seconds, setting a new standard for speed and quality in the field of 3D reconstruction!
Alongside this release, weāve also published a technical report that highlights how we achieve fast inference speeds with reduced baked illumination and material parameters.
š¾You can learn more and access the report here: https://stability.ai/news/introducing-stable-fast-3d
fluid jungle
SXSW Panel Proposal!
š£ Weāre thrilled to announce our panel proposal for SXSW featuring our CEO, Prem Akkaraju, and renowned journalist Kara Swisher.
š At the intersection of breakthrough and innovation, SXSW is a brilliant opportunity to share the importance of open models, the role of government, and the potential of AI to enhance human potential worldwide.
š³ļø Up to August 18th, the community will be able to review proposals, leave feedback, and vote for sessions that they believe would be a great addition to the conference.
You can cast your vote and secure our position here: http://panelpicker.sxsw.com/vote/153232
fluid jungle
James Cameron, Academy Award-Winning Filmmaker, Joins Stability AI Board of Directors
<@&1072166017565999114>
Today, our CEO, Prem Akkaraju announced that legendary filmmaker, technology innovator, and visual effects pioneer, James Cameron, has joined the Stability AI Board of Directors.
Cameronās addition represents a significant step forward in our mission to transform visual media. His artist-centric perspective, paired with his business and technical acumen, will be invaluable as we continue to build a full stack AI pipeline that unlocks new opportunities for creators to tell stories in ways once unimaginable.
Read more here: https://stability.ai/news/james-cameron-joins-stability-ai-board-of-directors
fluid jungle
Stable Diffusion 3.5 Launch!
@everyone
Introducing Stable Diffusion 3.5, our most powerful models yet.
This open release includes multiple variants that are highly customizable for their size, run on consumer hardware, and are free for both commercial and non-commercial use under the permissive Stability AI Community License.
You can download Stable Diffusion 3.5 Large and Stable Diffusion 3.5 Large Turbo from Hugging Face and the inference code on GitHub now. Stable Diffusion 3.5 Medium will be released on October 29th.
In June, we released Stable Diffusion 3 Medium, the first open release from the Stable Diffusion 3 series. This release didn't fully meet our standards or our communitiesā expectations. After listening to the feedback, instead of a quick fix, we took the time to further develop a version that advances our mission to transform visual media.
Our analysis shows that Stable Diffusion 3.5 Large leads the market in prompt adherence and rivals much larger models in image quality.
Stable Diffusion 3.5 Large Turbo offers some of the fastest inference times for its size, while remaining highly competitive in both image quality and prompt adherence, even when compared to non-distilled models of similar size.
Stable Diffusion 3.5 reflects our commitment to empower builders and creators with tools that are widely accessible, cutting-edge, and free for most use cases. We encourage the distribution and monetization of work across the entire pipeline - whether it's fine-tuning, LoRA, optimizations, applications, or artwork.
We canāt wait to see what you create!
fluid jungle
Stable Diffusion 3.5 Medium Release
@everyone
Stable Diffusion 3.5 Medium is here ā this open model is free for both commercial and non-commercial use. With 2.5 billion parameters, this model is designed to run āout of the boxā on consumer hardware, even on a toaster!
Whether a startup or creator, access to this technology shouldnāt be restricted by hardware limitations. We took a look at the hardware compatibility for running Stable Diffusion 3.5 Medium alongside other open-image base models. This model only requires 9.9 GB of VRAM (excluding text encoders) to unlock its full performance, making it highly accessible and compatible with most consumer GPUs.
This model delivers best-in-class image generation for its size, with advanced multi-resolution capabilities. It surpasses other medium-sized models with its prompt adherence and image quality, making it a top choice for efficient, high-quality performance.
Learn more on our blog here: https://stability.ai/news/introducing-stable-diffusion-3-5
scarlet grove
@everyone
ControlNets for Stable Diffusion 3.5 Large
We heard you want Control. š®
So, today we are adding new capabilities to Stable Diffusion 3.5 Large with the release of three ControlNets: Blur, Canny, and Depth.
The models are free for both commercial and non-commercial use under the permissive Stability AI Community License.
You can download the model weights now from HuggingFace and code from GitHub, with support in Comfy UI.
ā¶ļø Learn more on our blog here: https://stability.ai/news/sd3-5-large-controlnets
wide sluice
@here
Oh hey!
Apologies for the ping but I wanted to take a brief moment to introduce myself;
I'm Maxfield, current Chief Community Guy at Stability! I've been a Stable Diffusion user since November of 2022 and have been heavily involved in the AI media generation scene since. Some of you may know me from Civitai where I was also very community focused.
Lets be frank, The engagement from Stability has been lackluster lately. We have a lot cooking behind the scenes and we'd like to a better job of sharing what we're working on.
I'm hoping to immediately implement two initiatives aimed at changing that:
1. A feature request board where you can suggest and vote on models, tools, or workflows that would be beneficial to the community. What's the point of all this compute if we're not building stuff you want?
2. Getting the talented researchers, trainers and creators at stability to share more of their progress as they build the tools you love. People here are making some wild stuff, we shouldn't be the only ones to see it.
There's more planned coming down the pipe aimed at hobbyists, professionals, and those that want to become professionals, but this feels like a good starting point to kick us off on the right foot.
If you have any questions, thoughts or ideas it is literally my job (and pleasure) to listen, shoot me a message š
You'll be seeing me around, looking forward to hearing from you!
wide sluice
New Feature Request Board is Live!
Weāre making it easier than ever for you to share your ideas and help shape the future of Stability AI! You can now submit feature requests directly from Discord by typing /feedback or visit our new request board to submit ideas and vote on existing ones.
Your feedback helps us prioritize what we work on next, so let us know what youād love to see!
wide sluice
@here
š Get Featured on Stability AIās New Website!
Weāre inviting the Stable Diffusion community to showcase their best work in our Website Redesign Contest! Winning images will be featured on Stability AIās official website (with full credit) to highlight the incredible possibilities of Stable Diffusion 3.5.
āØ Creative Brief
šÆ Objective:
Weāre looking for images that feel fresh, impressive, and forward-thinkingāthe kind of visuals that make people stop and say, "Wow, that was made with AI?"
š Image Specs:
ā Aspect Ratio: 16:9
ā Fresh, and sleekāout of the ordinary
ā Inspiring, clean, and striking
ā Open-ended themeāas long as it conveysĀ innovation, beauty, and the future of creativity
āļø Technical Requirements:
- Must be created using Stable Diffusion 3.5 as a base
- Youāre free to use custom nodes, fine-tunes, or LoRAs!
ā ļø Guidelines ā What to Avoid:
š« IP-infringing content
š« Robots or apocalyptic themes (Weāre steering away from the ārobots take over the worldā narrative)
š« NSFW material, including blood or depictions of violence
šļø How to Enter:
- Create stunning AI-generated artwork using Stable Diffusion 3.5
- Submit as many entries as youād like ā no entry limits!
- Ensure all images are in 16:9 aspect ratio
- Submit your work here
š Who Can Participate?
This contest is open to US participants only (Sorry folks, legal requirement)
š Deadline
Submissions close on Friday, March 7th.
š Whatās in it for you?
š Your artwork featured on Stability AIās official website!
šØ Recognition & Community Showcase as part of our creative community
ā A third thing with the same meaning as the previous two! Look you get your stuff on the website okay? Thatās pretty dang cool.
We canāt wait to see your most inventive, boundary-pushing workāpush the limits, experiment, and showcase the future of AI creativity!
fluid jungle
Stable Audio Open Small šµ
Today weāre open-sourcing Stable Audio Open Small, a 341M-parameter text-to-audio model optimized to run entirely on Arm
CPUs.
This means 99% of smartphones can now generate music-production samples in seconds, right on-device with no internet required.
Built for fast, on-the-go creation, it turns your next quick idea into up to 11 seconds of audio. Generate drum loops, foley, riffs, and textures right where you are. No cords š just chords š¹
You can learn more here: https://bit.ly/3GUomcC
fluid jungle
Introducing Stability Seconds
Weāre kicking off Stability Seconds, a new series of mini tutorials designed for those who want to spend more time creating and less time configuring ā°
Each episode will break down how to use our generative tools into simple steps ā just enough to move you forward without slowing you down.
Today's topic: how to set up ComfyUI with our models, so you can start generating images using a workflow that doesnāt require an API key or GPU.
Now, itās your turn š https://www.comfy.org/