LTX 2.5 is out!!!! RUNE!! Are you making WF for LTX 2.5 already?
wow thats pretty fast!!!
LTX 2.5:
A medium close-up shot of the man with glasses and a beard in a blazer. Looking directly into the camera, he speaks the exact dialogue line: "I'm trying the new LTX 2.5, and this model do pull a lot of V-RAM compared to LTX 2.3, but quality is way better." His lips and mouth move accurately to match the spoken words. After finishing the sentence, he breaks into a warm smile and chuckles knowingly. Highly realistic facial expressions, stable background.
768x1024 10 seconds in less than 3 minutes..... lol RTX 5060ti
i changed the models node and it works already.... lol
Only thing ComfyUI needs to be on latest version to use LTX2.5. I had to upgrade to 0.32 as 0.31 was erroring out during prompt processing for me.
The workflows are practically identical, very minor differences.
But will update for 2.5 ; -) Its basically just the cfg node (not sure its even required, will test with the old and see if its any difference), and new sigma values
The workflows are practically identical, very minor differences.
But will update for 2.5 ; -) Its basically just the cfg node (not sure its even required, will test with the old and see if its any difference), and new sigma values
hey man, custom WF for 2.5 using custom audio, was working great, but sometimes it wont lipsynch what i type, you think you can make a IMAGE TO VIDEO simple CUSTOM AUDIO INPUT specially for LTX 2.5? MANY PEOPLE are saying its having issues with custom audio inputs on many pages.... and they are using other WF..... can you verify that?
Often a bit tricky with lip-sync. The model sometimes get confused and makes it a narrator instead.
What can help is transcribe what is said in the prompt.