Це відео не доступне.
Перепрошуємо.
FREE and Unlimited Text-To-Video AI is Here! 🙏 Full Tutorials (Easy/Med/Hard)
Вставка
- Опубліковано 11 чер 2023
- In this video, we look at 3 different methods to generate AI text-to-video easily. First, we look at Gen2, which is now publicly available and has free monthly credits. Then, I show you how to use open-source text-to-video models with Google Colab, and then how to install it locally.
Enjoy :)
Join My Newsletter for Regular AI Updates 👇🏼
www.matthewber...
Need AI Consulting? ✅
forwardfuture.ai/
Rent a GPU (MassedCompute) 🚀
bit.ly/matthew...
USE CODE "MatthewBerman" for 50% discount
My Links 🔗
👉🏻 Subscribe: / @matthew_berman
👉🏻 Twitter: / matthewberman
👉🏻 Discord: / discord
👉🏻 Patreon: / matthewberman
Media/Sponsorship Inquiries 📈
bit.ly/44TC45V
Links:
Colab - colab.research...
Gen2 - research.runwa...
Open Source GitHub - github.com/cam...
Cuda Checker - gist.github.co...
Local Install - gist.github.co...
CUDA Toolkit - developer.nvid...
Who ever is doing the text to video could save massive overhead by having a image generator make key frames or storyboards then running a "fill the gap" ai to smoothly transition between frames. Fun part is that platform would allow you to import storyboards and generate video from them.
Something Pixar for example is uniquely positioned to do given the storyboards and final products they could easily build an AI that makes full length movies.
That sounds amazing.
RunwayML has an image interpolation tool in addition to their Gen2 video generator. You provide a series of frames and it links them together. I have not used it yet but I’m positive it will save a lot of time for animation as you mentioned.
This is a very smart approach.
Great Video. The velocity of the improvements we are seeing is truly amazing.
Thanks again Marc!
I dunno how you ended up in my feed but I love your work! Keep it up!
Your insights are immensely valuable to me. I deeply appreciate your contribution and commend your exceptional work. Wishing you a fantastic day ahead!
The free OS version seems to be on a close level to the paid gen2.
Yes, and, it's iterating rapidly. New models are coming out that make the quality and length even better
Hi Matthew, Thanks for the video. I tried to follow the instructions, but it seems the project no longer works, the models have been removed from the /zeroscope_v1-1_320s file, is there an updated version of this project? here is the error I am getting
Error no file named pytorch_model.bin, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory /content/zeroscope_v1-1_320s.
No this vid just doesn't cut it unfortunately you shoot off do things come back past stuff in at lightning speed pop over here to get something to ad with this thing I put here and you say nothing about what you doing and there are no links to copy paste we have to read from your terminal. It's a mess. when i enter exactly the same code my output looks in d drive and c drive i have no idea why. errors everywhere.
Matthew, your Discord invite is impossible to accept. You have it set for verification but it just goes in a loop.
getting nerror as invalid number of channels in input image 'Vscn:conatins(scn)' where 'scn' is 24
i find the zeroscope > gen2 for Accuracy of text prompt to output
Sorry did you make the update video I can't find it
Have you experimented with running this on mac?
thanks man always updated because of you
You're welcome!
Thank you Matthew I appreciate you
thanks :)
What type of nvidia GPU do you have? because a 3060 is not the same as a 4090
I tried gen-2 for hours and only got static images with a zoom animation effect. That's all. I also told this in an email to them.
really?? that's surprising. it worked well for me.
@@matthew_berman Strange.
I got a mixed bag. It does something’s better than others. It does stock footage best. I think it’s interesting it really struggled with fantasy. I think that training data for dragons is limited to heavily copyrighted works. The kind of stuff you get in trouble for downloading 😂. It’s illegal to put DVD’s you payed for on your own computer, they are probably limited in what they can train with.
Its probably somewhere and im not seeing it.. can I use the local host on a 12g GPU? Or do i need to get a 16 and up?
Real impressive
"Billed yearly as $144"
No thanks.
No need, use the free and open source version I showed :)
Which gpu you have for this?
Awesome channel!
Amazing informative video. By any chance could you also share your gpu specs
Thank you! RTX 4090 24GB VRAM
sir is there any free source text to video tools on 5/06/2024 ?? please suggest some
Thankyou so much sir
hey mat is there any local video-to-video model?
oh good question. i'm not sure, but i'll look around!
@@matthew_berman 🙏Thank you🙏
Can use this model offline
Colab is fairly useless now unless you pay for it. Not a gamer but really need a GPU now for all this AI stuff to run local. I wonder if the latest Ryzen APU's can handle it?
Once again and as always
Major Alpha 💯 🧠
Quick question, can Anaconda help with Python code I'm already testing on Linux?
Thank you :)
Yes, Conda can help with existing code. I'm using it on every project now.
What is the minimum memory you need on GPU and vram? Thanks
Crazy how "easy" it is to be rich now but when I try to do anything it's literally undoable.
Very interesting
Would this work on a gtx 1650 max q?
Can you run it with automatic1111 webui?
13 seconds of laughter : ua-cam.com/users/shortsHtXVNxI6rig
it = iteration
thank you! that's what i thought
Gen 2 is a big disappointment why would I want two headed animals or person.
So imagine - we make an AI that takes these video frames and makes it more realistic, then we loop a few dozen times.
This was a quantum leap - we went from teams of people working for weeks using the greatest software fools made by mankind - To an AI system that makes video with nothing but Text input.
Your like the people who scoffed at the first transistor.
“Vacuum tubes are better”
@@spoonikle it's closed source and paid, if it's bad - it's bad
That’s the cool thing, i almost mendella effect a episode from a show, it’s kinda interesting
People watch dummer 💩
This aged like mold
do people care? i make a thoudsand websites and dont really use that much graphics stuff
Please do not say FREE in your heading when the software your advertising is not FREE.
Stop saying its FREE then tell us it has to be paid for. STOP IT.
so much dead ai crap on my PC now
a lot of you tech people gate keep so much either on purpose or on accident
ninnaman tunne unnake unlimited anta haktya
Dont say this is free. Dont lie to get views.
rubbish