Mega V10 (and lower) not working with Vantage I2V single model looper, V10 works as expected
#184
by
klaasb123
- opened
This is a bit of a cross post with this issue https://github.com/vantagewithai/VantageLongWanVideo/issues/9
Where I'm experiencing completely not relation to the start images. After some back an forth the suggestion was to test against v10 I2V (not mega) and there the same workflow does what it should do and follows the start images correctly (and second batch of 5 seconds etc).
I now expect the mega checkpoint is the cause of this but also not sure.
Some could it be that this is a bug in the merge of seperated I2V and T2V models?
"mega" is a T2V model with VACE added, so it works much differently than an I2V model. You can't use a typical I2V workflow with mega.