Credit where credit is due:
All credit to Ez4M for the unlimited length nodes ( https://civarchive.com/models/1499114/unlimited-length-i2v-framepack-lora-support?modelVersionId=1702158)
What makes this different from the I2V version:
You can go from text to video.
How it works:
I took the I2V version and replaced the input image with a T2V front end. This front end only generates 5 frames, then resamples them 3 times to get the quality up. Then it grabs the last frame and passes it off as the image input for the unlimited length nodes.
Limitations:
You can use loras to generate the first 5 frames. But the loras do not pass through to the video generation. So if you're looking to replicate a likeness, or get a certain pose or concept you can. However, if the lora is meant to guide the video from start to end, you will not get that.
*Ez4M's latest version does enable full video length loras. I have not attempted to perform this same modification on that version yet.
Description
First version. Does not do full video length loras.
FAQ
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.