Is it possible to use it with I2V ?
When I am trying to use this LoRA for ItV it works, but cropping the guides at start so first 33 frames are absent in the final video, making the video look like its starting from mid of animation.
Is there anyway to adapt this LoRA and its nodes for I2V pipeline so animation will start from the first input frame?
Want to say its best I`ve seen so far regarding face consistency for LTX, great work!
Ah never mind I forgot to connect Prompt Relay, the only problem left is that animation is feeling slow and usually ignoring last segment of the prompt.
Ah never mind I forgot to connect Prompt Relay, the only problem left is that animation is feeling slow and usually ignoring last segment of the prompt.
The slow motion actual happens while testing, and I found that, refining prompt may help.
Thanks, by testing a lot of different scenarios I figured out that the Prompt Relay were the biggest problem, so I just stick with LTXVImgToVideoConditionOnly for the first low res pass to make I2V properly work, and then use regular LTX prompting for both passes to make video action packed and it works!
Thanks, by testing a lot of different scenarios I figured out that the Prompt Relay were the biggest problem, so I just stick with LTXVImgToVideoConditionOnly for the first low res pass to make I2V properly work, and then use regular LTX prompting for both passes to make video action packed and it works!
Thank you for your feedback.