Dall-E, Midjourney, Stable Diffusion (etc) - who's playing, and with which?
Dall-E, Midjourney, Stable Diffusion (etc) - who's playing, and with which?
Out of context: Reply #793
- Started
- Last post
- 820 Responses
- PhanLo3
Not the best example but Open AI just released some demos of a new text to video generator called Sora that can do upto 60 seconds and is quite consistent.
-
https://openai.com/sora
-- Chris Darmon approves!palimpsest
- Very promising. No (or almost no) morphing/ jitter effects. When available?NBQ00
- https://youtu.be/SIt…jonny_quest_lives
- Seems like it might be a wee while, says they're still safety testing it. Some of the examples on that site are pretty mad.PhanLo
- wow, wow, wow the video samples on that link are bonkers good. I can't tell the difference from real video._niko
- "can't tell the difference from real video." really? every sample looks like 480p upscaled to 720p felt like i fell into the spike jonze 2001 filterjonny_quest_lives
- https://www.youtube.…jonny_quest_lives
- this already looks better than 99.9% of 3d shit out there. https://cdn.openai.c…_niko
- it still looks upscaled.feels like they took every HD LED TV demo video reel and trained their model on it but the outputhas less resolution and is more jitteryjonny_quest_lives
- https://www.youtube.…jonny_quest_lives
- i do like nightmare fuel though it is an aesthetic to itself: https://openai.com/s…jonny_quest_lives
- ^ slightly possessed demon grandma :-)PhanLo
- yeah creepy. I love this though, the stuff of dreams https://cdn.openai.c…_niko
- Def sounds like it'll be a while before it's public release. Seems like they are attempting to be ultra responsible to make sure its all PC and for all ages. ;)microkorg