331–340 of 437 posts
Earlier quoted context omitted.
Wait for the 4070, it should be around the perf of a 3080/maybe 4090 for ~500 bucks if rumors hold up. It is coming in a few months. NZXT used to make good pre-builts. Not sure which others have a good rep. DO NOT BUY DELL.
They were saying that before the 3080 was released. It will be cheaper and better than current hardware. Then when it came out you couldn't buy it for a year.
Earlier quoted context omitted.
https://huggingface.co/spaces/stabilityai/stable-diffusion Here you go.
the queue to make images was 1 now is 5
> You can join our dedicated community for Stable Diffusion here: [ https://discord.gg/stablediffusion ] Oh, discord... I've had so many problems trying to log into discord in the last couple of years that I've given up on it.
That's peculiar. I run a couple of big Discord servers and never heard anyone saying they have problems logging in.
Earlier quoted context omitted.
Yes, clone the repo ( https://github.com/CompVis/stable-diffusion ), download the weights and follow the readme for setting up a conda environment. I am presently doing so on my RTX 3080.
I can't believe i sold my rtx2070 last month, aaargh...!
Earlier quoted context omitted.
That's peculiar. I run a couple of big Discord servers and never heard anyone saying they have problems logging in.
First about 18 months ago they said I had an IP phone and could no longer log in (my carrier is Republic Wireless). When I contacted support and told them that I had been able to login with that number in the past they basically said "we're tightening up security, too bad so sad. get another phone number" Then recently I found that I was able to log in when I wanted to try midjourney (Republic changed something back…
Earlier quoted context omitted.
Not if you change the precision to float16. Should work on a smaller card. Tried on a 1080 with 8GB and it works well.
How would one do that? ----- Sorry my bad, found the answer. One simply adds the following flags to the StableDiffusionPipeline.from_pretrained call in the example: revision="fp16", torch_dtype=torch.float16 Found it in this blogpost: https://huggingface.co/blog/stable_diffusion mempko thank you for your hint! I was about to drop a not insignificant amount of money on a new GPU. What does one lose by using float16 re…
Edit: Just to be clear, your intuition that it could cause issues is certainly merited - and not _all_ models can be trivially converted from fp32 to fp16 without some new error accumulating (during inference). Variational autoencoders like VQGAN and GAN's are particularly prone to such issues.
But in this case, it's all upside.
Earlier quoted context omitted.
Marvel Studios makes billions of dollars every year
And I'm fairly confident in saying that AI will never be able to generate a Marvel movie! (Not in our lifetimes, anyway.)
Earlier quoted context omitted.
And I'm fairly confident in saying that AI will never be able to generate a Marvel movie! (Not in our lifetimes, anyway.)
Considering the phenomenal progression of Dall-E-1 to Dall-E-2 in just over a year, I'm not really understanding your confidence on the limits of AI content generation.
* Coherent video
* Characters with backstory
* Dialogue (including jokes and witty banter)
* Music
...among many other things. Plus, the training set for video is orders of magnitude smaller than for digital art. (And is additionally burdened with copyright issues.)
As I see it, there's simply no path from the DALL-E of today to something like that. And all for art that, essentially, "says nothing and means nothing".
I've been looking forward to this. The license however strikes me as too aspirational, and it may be hard to enforce legally: > You agree not to use the Model or Derivatives of the Model: > - In any way that violates any applicable national, federal, state, local or international law or regulation; > - For the purpose of exploiting, harming or attempting to exploit or harm minors in any way; > - To generate or dissem…
The most interesting part, to me, of a release like this is the amount of "please don't abuse this technology" pleading. No licence will ever stop people from doing things that the licence says they can't. There will always be someone who digs into the internals and makes a version that does not respect your hopes and dreams. It's going to be bad. As I see it, within a couple years this tech will be so widespread and…