You got my sub... Idk weather this will work for me or not when I'm trying... But the way you presented was appreciated... I will look your other videos also ✌💯
AI tools are transforming how we live and work, offering unparalleled efficiency and innovation. With ssmythOS, managing tasks and workflows has never been easier. #AITools #Innovation #ssmythOS
A free AI tool that can control expressions and emotions in videos? That's incredible! Can't wait to give Live Portrait on Google Colab a try. Has anyone tested it yet?
Thank you for the information. I would like to ask: what is the Colab time-limit to process image?. HuggingFace only support short videos < 10 seconds.
So i tried this and while it does work, it's kind of lame it doesn't forward the audio unless i am missing something. It's also limited in range of what it can do, if a subject is moving at all in the frame of the original video, the portrait stays in place and looks really strange. so source has to be basically a cropped and nearly completely still face with only slight movements. Some facial expressions don't work at all, at least in an example i tried. It's a cool concept but again another AI tool you play with a couple minutes then seen all you need to see with the tool. but content creators may find use of it.
Can you modify the code on colab notebook to accept a source video (instead of a source image)? :) the model can also take a source video but it requires some modification. Thanks
You need to keep 1:1 ratio, fix the camera in one place, sit calmly and only focus on facial animations. Holding camera in hands will result in shaking
Google Colab was working perfectly well till now. But now no files comes in Animations tab after the entire processing. Is the link broken? Is anyone facing same issue and can please update the Colab link or suggest a solution.
Find an original zip, it weighs 5gb as far as i remember, the tree of zips has multiple versions from different dates, it has a shell script that opens a local server that has img to vid and vid to vid, but you need a beefy gpu because vid to vid 30 frames fullhd takes about 1 minute on 4080 super but it depends on how tricky you can get things as efficient as possible
Yes truely , shame on them really , i spent days to install it locally and when i failed and tried it with google colab , the results disapointed me . I think the errors are intentional by the developer , because of the frequent updates and changes that i had noticed in github
hugging face requires me to use a square video and a better gpu. Replicate wants me to pay with every generation, colab made my image looks so weird in the video.
this AI image to video tool sounds absolutely amazing. The ability to control expressions and emotions is such a game-changer for creative projects.
fake
You got my sub... Idk weather this will work for me or not when I'm trying... But the way you presented was appreciated... I will look your other videos also ✌💯
what weather are you expecting, rain or sun or snow?
Thank you! All the other videos on this only showed the complicated methods!
WOW! great find!
Excellent presentation of information, you are top mate :) !
Great tools. Very good video. Thank you
wonderful and informative content brother
Thank you for your efforts !!
AI tools are transforming how we live and work, offering unparalleled efficiency and innovation. With ssmythOS, managing tasks and workflows has never been easier. #AITools #Innovation #ssmythOS
A free AI tool that can control expressions and emotions in videos? That's incredible! Can't wait to give Live Portrait on Google Colab a try. Has anyone tested it yet?
yeh not free
@@Surya_.Choudhary yes it is! Im running it locally for free as we speak
Thankyouuuuuuuuuuuuuu so much🥰🥰🥰🥰
😢
Thank you for the information. I would like to ask: what is the Colab time-limit to process image?. HuggingFace only support short videos < 10 seconds.
Subscribed ❤
I experimented with this, it's AWESOME! I only wish it was able to animate FULL BODY, or at least face, hands and upper body 😥
i have one and its free
@@TRADDLOGICwhats the name?
Whats the name?@TRADDLOGIC
Hello, can this tool be controlled in real time through the camera?
Comfy ui is the best way to do this atm
When I use Colab and generate the video, why there is no animations folder under LivePortrait folder? I couldn't find the output video..
Its strange, make sure that you copied the path on correctly
@@planetai217 I used the correct path. I changed the name of the image file, now it's working.
May I ask what was the previous nane and what did you change it to? @@cody.L881
@@cody.L881 me too
So i tried this and while it does work, it's kind of lame it doesn't forward the audio unless i am missing something. It's also limited in range of what it can do, if a subject is moving at all in the frame of the original video, the portrait stays in place and looks really strange. so source has to be basically a cropped and nearly completely still face with only slight movements. Some facial expressions don't work at all, at least in an example i tried. It's a cool concept but again another AI tool you play with a couple minutes then seen all you need to see with the tool. but content creators may find use of it.
Thanks brother 👍
do you know any ai local software or way in local pc that i give image avatar + audio file and it give me video of my avatar speaker?
error: The requested GPU duration (240s) is larger than the maximum allowed retry in -1 day, 23:59:59.. what does that mean? can anyone help?
Same
Nice one . How can I get realtime lip synch avatar for chat bot ?
good video!
why in replica i cannot upload a photo from my computer and there is not a run button?
Kling 🔥 🔥 🔥
When I upload video it shows me an error like video In square ratio
Can it work on animals facial features?
You just got a subscriber
😊
Hugging face doesn't work
Google collab (other comments say it doesn't work)
so basically this just doesn't work.
Find the official zip and you can run it locally on your pc, only if you have a good rig
Can you modify the code on colab notebook to accept a source video (instead of a source image)? :) the model can also take a source video but it requires some modification. Thanks
Run it in comfy UI on your computer
39C weather at your home! wow that's hot
Does it work with only indian accent or any English? 😁
Nice video
Can you run it on Kaggle instead of Google Cola?
why my end results end up being jiggling!?? it's crazy shaky!
You need to keep 1:1 ratio, fix the camera in one place, sit calmly and only focus on facial animations. Holding camera in hands will result in shaking
Google Colab was working perfectly well till now. But now no files comes in Animations tab after the entire processing. Is the link broken? Is anyone facing same issue and can please update the Colab link or suggest a solution.
Find an original zip, it weighs 5gb as far as i remember, the tree of zips has multiple versions from different dates, it has a shell script that opens a local server that has img to vid and vid to vid, but you need a beefy gpu because vid to vid 30 frames fullhd takes about 1 minute on 4080 super but it depends on how tricky you can get things as efficient as possible
this really looks like a good tool, I think I will use this in the wrong err I mean in the right way. lol
- where can I find such short references with emotions??
You can record yourself
@@planetai217
You talk too fast. With your accent as strong as it is, it's hard to understand for most English speakers.
I understood it fine.
Can i use it in mobile
is this real time?
But A.I.video will not monetized by youtube.
I still don't know how to get more views and subscribers :( I read many these type of videos already
There is a limit, next generations is paid
unfortunately not a single option is free
I want to use by install. Can you help.?
you can see the install guide at Ai Search channel
Einstein uncle 😂
😆
Plz what os thé maximum duration of one vidéo ?
On local - practically any duration
works much worse than in demo advertising!!
Yes truely , shame on them really , i spent days to install it locally and when i failed and tried it with google colab , the results disapointed me . I think the errors are intentional by the developer , because of the frequent updates and changes that i had noticed in github
Of cource! Miracles dont exist yet!
If everything's done correctly, it will be good
hugging face requires me to use a square video and a better gpu. Replicate wants me to pay with every generation, colab made my image looks so weird in the video.
not work
can it generate landscape videos as well ?
thank you , but it takes too long !now its about 10000 secs that passed and animation didnt creaated!
3 months 🥵🥵🥵🥵
Can you put an English subtitles on what you're saying
Kling ai
Why is there no sound in the generated video?