Show HN: Stable Diffusion v2 web interface

https://getimg.ai/text-to-image

by jkbl • 4 years ago

72 44 4 years ago

getimg.ai

mlu509• 4 years ago

Wow. That’s pretty fast!

devxpy2• 4 years ago

Shameless plug - We don’t require logging in!

https://app.gooey.ai/CompareText2Img/?example_id=1uONp1IBt0Y

Traubenfuchs3• 4 years ago

But do you allow high res nsfw Emma Watson prompts?

rejectfinite• 4 years ago

I can't wait for the morally concerned to rail against Stable diffusion once they find the /b/ threads and AI is going to get locked down... Its going to happen. "This AI tool enables sick PREDATORS!!"

Aeolun2• 4 years ago

Doesn’t that just get you those completely black images?

mrkeen• 4 years ago

Does it generate images though? I gave it a prompt around an hour ago and it's still just spinning.

devxpy2• 4 years ago

Yeah, seems like its been hugged to death a bit. A single GPU is not quite enough.

Completes in around a minute for me right now

kburman• 4 years ago

I hate it when website hide login popup until user click on action button. It's such a dark pattern.

ftufek• 4 years ago

FWIW, you're not the only one. On our product, we experimented adding login when the user tries to use any of our tools and something like ~60-70% of users dropped off. So essentially, whoever implements that is killing their user funnel at the very beginning which I'm guessing will have larger effects down the road (fewer users experience "magical moments" which leads to less word-of-mouth, etc).

On the flip side though, I'm sure it may help metrics in the short term especially since it's expensive to offer a service like this (which requires expensive GPU servers).

_Microft• 4 years ago

Requires login.

Saying that it requires login before someone spends time to think up a description for an image would have been nice (because I won't sign-in to try something anyways).

Edit: rephrased to remove overly negative language.

jkbl7• 4 years ago

I don't see this as baiting. The link to a tool page was shared. It's transparent on the homepage and in the header. The text to image page just shows how it works without the need to create an account.

Traubenfuchs3• 4 years ago

It's absolutely not transparent. Showing an active "start process" button that leads to a sign up / login form isntead of starting the process is absolutely not what users want or expect.

version_five• 4 years ago

No, it's definitely a dark pattern. Everything about the page suggests that pressing the "generate 4 images" button will generate four images. And then it doesn't and tries to lever that anticipation to get a sign-in. It ends up just making users angry.

If you need a sign-in, e.g. to prevent abuse, you could start by explaining that a sign-in is required and why. At least then people won't be angry when they are ambushed by it

Aeolun2• 4 years ago

I think this is the very definition of baiting.

davidkunz• 4 years ago

This site doesn't require a login: https://tinybots.net/artbot?model=stable_diffusion_2.0

Posted by @davely today.

ShamelessC• 4 years ago

Proof that people just upvote stuff with stable diffusion in the title without clicking the link or taking more than a cursory glance.

88stacks• 4 years ago

Here is another that is free AND allows you to create your own custom model: https://88stacks.com yes it does require a login

IceHegel• 4 years ago

Wow v2 seems worse than v1.5.

kyleyeats• 4 years ago

They took out a bunch of artist references, and a bunch of celebrities. The two things that made SD stand out. It's no longer the fun one.

frobolo• 4 years ago

Agreed. I like making abstract stuff, much more fun to be had with v1.5.

lumenwrites2• 4 years ago

Great project!

I have a question - how can you afford to run this? I looked at the server costs on google compute and runpod.io, and the ones with GPUs powerful enough to run SD are pretty expensive. I'm guessing you don't have a lot of paying users at the beginning, so I don't think you have enough revenue to cover the costs (at least when you're just getting started).

Do you just spend a lot of money to fund this, and hope that once you have enough users it will pay for itself? Or did you find some affordable way to run a server? If yes, and if it's not a secret - where are you renting the server, what are its specs, and how much does it cost?

permanent• 4 years ago

I'd love to know this as well.

nacs• 4 years ago

Stable diffusion only needs 4GB of VRAM to run on the low end so you can rent low-end consumer GPUs (nvidia RTX for example) for around $0.10 an hour to do the renders.

jkbl7• 4 years ago

Yeah, but that's only true when you use one model for yourself. More VRAM is needed for running such a service. It currently loads 6 models per single GPU. And I think I have some VRAM left to add even more.

UncleEntity• 4 years ago

I was running it for free on google colab here and there with no problems.

Haven’t tried the 2.0 out yet so not sure how that’s going.

jkbl7• 4 years ago

At the beginning it was just burning money, but it stars to turn into profit.

The only secret to making it affordable is to autoscale based on a current image generation traffic. It runs on a mix of Tesla V100 and RTX 3090 from runpod.io vast.ai and Lambda Labs.

I actually went through a couple of iterations with it, starting as completely free service, then offering one time payments for image credits and now settled on subscriptions as I see there is a demand for the product. Especially AI Editor[1] which I think offers a unique value to the users.

[1] https://getimg.ai/editor

lumenwrites2• 4 years ago

That's amazing, thank you for sharing!

If not a secret, how are you marketing it?

synapticpaint• 4 years ago

On AWS, a g5 instance costs $1/hr. I can generate roughly 10 images per minute (should be able to get this down with some optimization), so 600 images per hour, so the cost per image is 1/6 of a cent, before adding overhead (idle time, start up/shut down).

Source: I run a similar project: https://synapticpaint.com/

I also offer dreambooth model training for around $2-$4 / model as well as inference on custom dreambooth models. Inference on custom models is where things get a little tricky because if users are using different models and you're loading up new models all the time just to generate 6 images, then that quickly becomes the majority of the work load, drastically pushing up the inference cost. I haven't solved this problem yet. If you have any great ideas, feel free to email me (email in profile)!

raymond_goo• 4 years ago

https://vast.ai/ (not affiliated)

mechanical_bear• 4 years ago

Verify email? No thanks.

russellbeattie• 4 years ago

Exactly. Put that bullshit up front before I waste time entering a prompt so I know not to bother.

Hoefner• 4 years ago

I can’t even register with my own email domain because of “We do not support disposable email addresses!” error

jkbl7• 4 years ago

It's a measure to prevent multi accounts. Some people really overuse free offering. Either your domain/IP is on a spam list or your email server may be misconfigured.

NVQXE23I2• 4 years ago

How is this a measure? I have my own domain, so why would I make several accounts for your service? I just need one.

And labeling a domain as disposable, when it's clearly isn't, sounds a bit strange to me.

isuleman• 4 years ago

Well you can also block domains from fakemail.net to prevent disposable emails.

tuhins• 4 years ago

cool, great to see more and more of these coming up!

you can also check out our demo for Stable Diffusion v2. it's free to use, with no email sign-in required and no credit system.

https://app.baseten.co/apps/VBlnMVP/operator_views/nBrd8zP

LegitShady• 4 years ago

Personally, having played around with both for only an hour with identical prompts, v1.5 seems to get better results.

joe_the_user• 4 years ago

It seems like v2 is more photorealistic but it more often fails to follow any complex prompt.

thundergolfer• 4 years ago

If you prefer interacting with v2 via code, Erik Bernhardsson got v2 running on https://modal.com in 20 minutes, and it'd take 5 minutes to copy and run his Github gist: https://twitter.com/bernhardsson/status/1595960120021696512.

---

To OP, I reckon you should allow 1 generation before hitting us with the login prompt. May lift conversion.

NVQXE23I2• 4 years ago

Came here to point out that my personal email is wrongfully listed as a disposable email address. It must be that it's a three letter url with a .nl TLD.

Maybe take a look at your mail checking API?

Show HN: Stable Diffusion v2 web interface

Related Stories