New Mars Forums

tahanson43206

Moderator

Registered: 2018-04-27

Posts: 19,421

Sora - New service from OpenAI - video from text

NewMars members have been living in a text only environment for 20+ years.

In recent times, some members have shown how to use drawing and graphics tools to help to illustrate their ideas.

Now it appears our members may have a new (powerful) tool to help to show the ideas they've shown us in words, now in video.

This topic is offered for members to show their creativity.

imgur.com will hold video, and there may be other providers available.

Video generation models as world simulators
We explore large-scale training of generative models on video data. Specifically, we train text-conditional diffusion models jointly on videos and images of variable durations, resolutions and aspect ratios. We leverage a transformer architecture that operates on spacetime patches of video and image latent codes. Our largest model, Sora, is capable of generating a minute of high fidelity video. Our results suggest that scaling video generation models is a promising path towards building general purpose simulators of the physical world.

Above is from a web page about Sora available from OpenAI.com.

Research index
Search the database

Search

Filter and sort
Research Papers
Feb 15, 2024February 15, 2024
Video generation models as world simulators
Video generation
Research

Above is from another page on the Sora web site of OpenAI.com

(th)

tahanson43206

Moderator

Registered: 2018-04-27

Posts: 19,421

Re: Sora - New service from OpenAI - video from text

This post is reserved for an index to content our members may contribute over time.

(th)

tahanson43206

Moderator

Registered: 2018-04-27

Posts: 19,421

Re: Sora - New service from OpenAI - video from text

As a follow up to the new Sora topic, here is a screen scrape of a page about developer opportunities using OpenAI tools:

Playground
Assistants
Fine-tuning
API keys

Apache Internal Server Error objected to a string it found in the rest of the quote.

(th)

If anyone is interested, here is a link to the developer page:
https://platform.openai.com/docs/overview

(th)

Mars_B4_Moon

Member

Registered: 2006-03-23

Posts: 9,776

Re: Sora - New service from OpenAI - video from text

There are so many new developments in AI text to video art creation services now

Mars_B4_Moon

Member

Registered: 2006-03-23

Posts: 9,776

Re: Sora - New service from OpenAI - video from text

AI has made great videos I have seen some of them but I have limited my time with these chatbots and painting artificial creation.

I'm stepping back from the cyber machine a bit I am not against it but will be happy to support those who like gaming and AI and all that artificial social media and tech image creation, the Internet in a ways can have people hooked inside their own world, similar in ways to a false interaction, a video game behavioral addiction almost like an addiction disorder however it can have great use as a tool but I think real life creativity and real life human interaction is also important.

Text to video just shows how fast it is all moving today.

This was 2 months ago

a video

'Introducing Lumiere: A space-time diffusion model for video generation'
https://www.youtube.com/watch?v=f9ThAzZs32M

Every few weeks you hear another spin off from OpenAI becoming its own thing, Sora seems to have been around a while coming from works on other text-to-image models, .

what is next, the sensation, 3d vibration and audio sensation, almost a 'Holodeck' the ball globe hologram and Sphere (venue) of Las Vegas where you play virtual gold or live out a simulation on Mars missions

https://www.youtube.com/watch?v=taOG0O2MYrU.

many examples here also
https://twitter.com/billpeeb

and

'Create an entire video from a single prompt'
https://abcaishow.com/

clip-guided-diffusion
https://replicate.com/afiaka87/clip-guided-diffusion

You maybe need to know Spanish but maybe you can get an online translator

https://colab.research.google.com/drive … STU50N9EeT

Maybe Chinese, an English-to-simplified Chinese translator icon that appears after 9 characters are typed.
https://agc.platform.baai.ac.cn/CogView/index.html

'Text-to-Image Summary'
https://softologyblog.wordpress.com/202 … e-summary/

'dribnet'
https://replicate.com/dribnet

"Dark Future" Short
https://youtu.be/gU64SPk91FI

Religion, Goats, Horses, and the Apocalypse imagery through AI ' Zeroscope '
https://www.youtube.com/watch?v=RPbpFEBeNNY

video game style "Street Runner" by Zeroscope with sounds
https://www.youtube.com/watch?v=gLdFf-4a51Y

Activating humanity's potential through generative AI
https://stability.ai/
Open models in every modality, for everyone, everywhere

It seems to be Born out of the OpenAI by Ilya Sutskever, Greg Brockman, Trevor Blackwell, Reid Hoffman, Jessica Livingston, Peter Thiel, Amazon, the India multi national Infosys, YC Research, Canadian Yoshua Bengio, Microsoft,  Vicki Cheung, Andrej Karpathy, Facebook or Google or You Tubes, the British Deepmind, Durk Kingma, Jessica Livingston, John Schulman, Pamela Vagata, and Wojciech Zaremba, with Sam Altman and Elon Musk serving as the initial Board of Directors memberssomeone posted a site which linked to a video some weeks back, it might have been you Tahanson or SpaceNut maybe but 'Sora' would be a common term or name in Japanese or Korean, there were also links on bad 'customer-service'

'Introducing Open AI's Sora: Text to Video'

https://www.youtube.com/watch?v=hRzzcLvsG8o

Last edited by Mars_B4_Moon (2024-04-06 10:38:07)