Any plans to support the first released (Skyreels)I2V model based on Hunyuan?

[shanque56666]

https://github.com/SkyworkAI/SkyReels-V1

SkyReels-V1 I2V model is open for several days, is there any plan to support lora-training of it ?

[sdbds]

When I have time, I will modify the code.

[kohya-ss]

I'm interested too. I'd like to start working on at least the inference soon. I'll post here if I start working on it so as not to overlap with sdbds's work.

[sdbds]

I'm interested too. I'd like to start working on at least the inference soon. I'll post here if I start working on it so as not to overlap with sdbds's work.

Based on my observation, the main difference seems to be that I2V only modified in_channel from 16 to 32.
I just need to change in_channel to the args parameter