Hi, I need help, I followed your tutorial but I get an error when generating video in the UI. I get AssertionError: config must have _class_name=CausalVideoAutoencoder What could be the cause ? THX
To fix the "AssertionError: config must have _class_name=CausalVideoAutoencoder" do "git checkout a5ab70c" on your LTX-Video folder. This will revert to an older version where they haven't changed the class name in .../vae/config.json among other things. This model does not seem to work on GTX 1070 (runs out of vram), maybe it only works for RTX cards.
I need someone to confirm that solution, I'm scared to run that command, it might mess up worse, and I'll have to download 120GB again lol. I searched for that error on Google and found something related to the program's developer, maybe there's a solution there too. I have no idea.
@@nvadito7579 I tried it but got this "(LTXVEnv) D:\AI\LTX Image To Video AI\LTX-Video-gradio-ui\ltx_video>git checkout a5ab70c fatal: Unable to create 'D:/AI/LTX Image To Video AI/LTX-Video-gradio-ui/.git/index.lock': File exists. Another git process seems to be running in this repository, e.g. an editor opened by 'git commit'. Please make sure all processes are terminated then try again. If it still fails, a git process may have crashed in this repository earlier: remove the file manually to continue."
@@nvadito7579 also tried deleting index.lock which is supposed to stop programm from changing updates and got this " error: pathspec 'a5ab70c' did not match any file(s) known to git" so apperantly there already is a newer version than a5ab70c or whatever
Thanks for your input, but like another person here I got "error: pathspec 'a5ab70c' did not match any file(s) known to git". Working with other AI text and image apps, I know this open source stuff changes names and versions almost daily. It's hard to keep instructions current and correct!
Hello great project but I was wondering how much gb does it take total ? the lightricks took 80 gb and i still have to install pixart-alphae but i only have 57gb left, is it going to be enough ? Thanks !
Can you share the error, I am trying to make the project work on the latest version of LTX video from comfy UI. Will upload a new video instead of keeping this.
Hey, great tutorial, but I get the following error: "AssertionError: config must have _class_name=CausalVideoAutoencoder" . Any help on fixing this? Thanks.
Still it should work, without it I tried installing triton in this video at 3:50 seconds. The triton requirement should be in the video description, there is a GitHub link th-cam.com/video/Fk6JQ8fufyk/w-d-xo.html I will make a refresh on LTX Video
I can get things to work on generating an error: AssertionError: config must have _class_name=CausalVideoAutoencoder. I have no experience in programming... GPT takes me on this to the Illusiana swamp :(...
It should work on RAM after making changes in the code, it will be a trial and error if I can manage spending time on it. Right now got caught up in other work.
Nice Video, just wanted to know if it can run in 6Gb of vram (image to video) UPDATE : yes it works on RTX 3050 6gb laptop and 16gb of ram and an i5 13th gen so u really can do stuff with it.
The switch for low vram if you saw in the video. Make sure to toggle it On. Yeh, it should not blow up. It would be terrible, this is the first time i did something with AI and python, trying to optimize some other project as well for low vram. I may start questing my self if bad happns
@@Aininja94 Update it works normally (a side from the usual gradio bugs and some times it crashes) after switching that toggle and yeah it is very impressive, thx for ur support and the tutorial I really appreciate that !
Add me to the list getting "AssertionError: config must have _class_name=CausalVideoAutoencoder". I followed your instructions precisely, twice, downloading 120GB only to throw it away and do it again for the same result. Your effort is appreciated, of course, but if it becomes obsolete due to change (which I know is very rapid in this field), it's only responsible to update it or (less preferable) remove it. Thank you.
I checked, there are multiple changes by LTX team in the lightricks huggingface files which is causing the error. I have updated the pined comments for others and trying to understand the changes for a fix
![แค่บอก - Atom ชนกันต์ [OFFICIAL MV]](http://i.ytimg.com/vi/kgwhtg8bAxY/mqdefault.jpg)
Clone lightricks older commit from hugging face. New update is giving error, need to be fixed.
Thanks for the video, any updates regarding the CausalVideoAutoencoder issue?
Hi, I need help, I followed your tutorial but I get an error when generating video in the UI.
I get AssertionError: config must have _class_name=CausalVideoAutoencoder
What could be the cause ? THX
Was busy in a video just uploaded, I got two comments with the same error, i will try and see it the error happens and let you know the fix
Same here.
please do more videos like this. I'm always following your videos. thanks for share this video with us
Thanks for the support, I got busy in work. Will get back to videos from today
Ty ❤
Man, I can't wait to see how much this will evolve in the next year.
If you knew...
Can you walk through how you installed Anaconda for this please?
To fix the "AssertionError: config must have _class_name=CausalVideoAutoencoder" do "git checkout a5ab70c" on your LTX-Video folder. This will revert to an older version where they haven't changed the class name in .../vae/config.json among other things.
This model does not seem to work on GTX 1070 (runs out of vram), maybe it only works for RTX cards.
I will check on the error and update. It should work, I think it's the error. Some packages might have updated
I need someone to confirm that solution, I'm scared to run that command, it might mess up worse, and I'll have to download 120GB again lol.
I searched for that error on Google and found something related to the program's developer, maybe there's a solution there too. I have no idea.
@@nvadito7579 I tried it but got this "(LTXVEnv) D:\AI\LTX Image To Video AI\LTX-Video-gradio-ui\ltx_video>git checkout a5ab70c
fatal: Unable to create 'D:/AI/LTX Image To Video AI/LTX-Video-gradio-ui/.git/index.lock': File exists.
Another git process seems to be running in this repository, e.g.
an editor opened by 'git commit'. Please make sure all processes
are terminated then try again. If it still fails, a git process
may have crashed in this repository earlier:
remove the file manually to continue."
@@nvadito7579 also tried deleting index.lock which is supposed to stop programm from changing updates and got this "
error: pathspec 'a5ab70c' did not match any file(s) known to git" so apperantly there already is a newer version than a5ab70c or whatever
Thanks for your input, but like another person here I got "error: pathspec 'a5ab70c' did not match any file(s) known to git". Working with other AI text and image apps, I know this open source stuff changes names and versions almost daily. It's hard to keep instructions current and correct!
Hello great project but I was wondering how much gb does it take total ? the lightricks took 80 gb and i still have to install pixart-alphae but i only have 57gb left, is it going to be enough ? Thanks !
The other was around 50gb if I remember
Installed yesterday (6-7 hours😒) but "error" occurs in preview window when it start to generate 😒😒
Can you share the error, I am trying to make the project work on the latest version of LTX video from comfy UI. Will upload a new video instead of keeping this.
@@Aininja94same "AssertionError" mentioned by other people in the previous posts
Will It run of 4gb Vram NVIDIA GeForce GTX 1650, with 16gb ram
I don't think it will, minimum 6gb it should be.
Did you tried flux gguf model? For images?
Hey, great tutorial, but I get the following error: "AssertionError: config must have _class_name=CausalVideoAutoencoder" .
Any help on fixing this? Thanks.
Was busy in a video just uploaded, I got two comments with the same error, i will try and see it the error happens and let you know the fix
i got this error : import error: No module named 'triton'
pls help me
Still it should work, without it
I tried installing triton in this video at 3:50 seconds. The triton requirement should be in the video description, there is a GitHub link
th-cam.com/video/Fk6JQ8fufyk/w-d-xo.html
I will make a refresh on LTX Video
@@Aininja94 when i click generate video but error
I got error message in generated Ouput window..what should I do?
I will make a updated video, ltx team had updated the code
@Aininja94 thanks
Does it work for AMD?
Sorry it won't
bro next time i want to run it does i need to re download all the process? or do only a short thing?
You have to download for the first time only
@Aininja94 but i cant do it again can tell me what need to do
u have Instagram or any social media? @@Aininja94
I am not active on social media but you can msg on Instagram, you should find by searching ai_ninja94
Any chance there is script for google colab? I am not free tier so I just need to know how to script it for it thank s
Hmm it's not on colab. I think I have to make one.
Please create video @@Aininja94
can you make a GGUF for LMStudio? TIA
Can you change the style of video like cartoon, cyberpunk style, etc?
It has to be mentioned in the prompt
I can get things to work on generating an error: AssertionError: config must have _class_name=CausalVideoAutoencoder. I have no experience in programming... GPT takes me on this to the Illusiana swamp :(...
What are the hardware specs needed?
In the video I am using a 10year old computer with 16gb ram and 4060ti 16gb graphic card.
If you see the video. It is using around 6gb of vram
where is the requirements.txt file ? it does not appear
After cloning the project you should have a requirements.txt file
It's there in the repository
is it also work in intel processor
Did you tried on a system without gpu?
@@Aininja94 yes
@@Aininja94 yes, it shows error.
It should work on RAM after making changes in the code, it will be a trial and error if I can manage spending time on it. Right now got caught up in other work.
bro how many gb is the filtering content download?
Didn't get by the filtering content, the project took around 60gb space.
Nice Video, just wanted to know if it can run in 6Gb of vram (image to video)
UPDATE : yes it works on RTX 3050 6gb laptop and 16gb of ram and an i5 13th gen so u really can do stuff with it.
It should, I tried limiting the my vram for testing but it's not happening.
I am waiting for someone commenting it worked 😅
@@Aininja94 oh ok will try and comment back, hope it doesn't blow up 😂
The switch for low vram if you saw in the video. Make sure to toggle it On.
Yeh, it should not blow up. It would be terrible, this is the first time i did something with AI and python, trying to optimize some other project as well for low vram. I may start questing my self if bad happns
@@Aininja94 Update it works normally (a side from the usual gradio bugs and some times it crashes) after switching that toggle and yeah it is very impressive, thx for ur support and the tutorial I really appreciate that !
Great, it works on 6gb then.
Please make a tutorial of mimic motion
It's there on the channel
th-cam.com/video/2P49SCJYEbY/w-d-xo.html
@@Aininja94 thanks 👍
But i need mimic motion which is used to video to video influencer reels copy
I don't know if mimic motion made video to video, do you have the project link? I will check it out
What about image to video ?
I showed two example of image to video at 6:23
Add me to the list getting "AssertionError: config must have _class_name=CausalVideoAutoencoder". I followed your instructions precisely, twice, downloading 120GB only to throw it away and do it again for the same result. Your effort is appreciated, of course, but if it becomes obsolete due to change (which I know is very rapid in this field), it's only responsible to update it or (less preferable) remove it. Thank you.
I checked, there are multiple changes by LTX team in the lightricks huggingface files which is causing the error.
I have updated the pined comments for others and trying to understand the changes for a fix