Hugging Face's logo

larryvrh
/
MiniMax-H3-Turbo-Lora

Text-to-Video
MiniMax H3
text-to-audio
audio-video
lora
comfyui
Model card Files
xet
 Community
35

Does it support ref2v?

#10
by Pevernow - opened
Discussion
Pevernow

If not, are there any subsequent LoRa training plans for ref2v?
larryvrh
 Owner

Not yet but planned.
SekiroRong

FL2va turbo lora seems work well on ref2va
rocky533

It work the same on both from what i can tell
bradac
•
edited 7 days ago

It's breaking identity reference on my tests. The first checkpoint at around 14 steps seems to be the least broken for Ref2V.
winterSoul

It doesn't work when referencing a video in ref2v.

ComfyUI Error Report

Error Details

  • Node ID: 214
  • Node Type: MiniMaxH3TurboLoRA
  • Exception Type: AttributeError
  • Exception Message: AttributeError: 'MiniMaxH3Model' object has no attribute 'diffusion_model'
pmoras

I don't know if my issue is the same as @winterSoul , but I've also noticed an issue when trying to run the turbo lora with an audio reference.
It appears to be an issue with the Turbo LoRA load node though, and an issue is already on the GitHub repo: https://github.com/Larryvrh/ComfyUI-MiniMax-H3-Turbo/issues/9

Sign up or log in to comment