Finetuning I2V Models

[mertalbaba]

Hello, do you have any full fine-tuning or LoRA fine-tuning codes for I2V models such as CogVideo5B-I2V or HunyuanVideo? Basically I am looking for text + image conditioning during fine-tuning but do you support this?

[YingqingHe]

Yes, we currently support finetuning of CogVideo5B-I2V with text + image control. You can check the documentation here: https://github.com/VideoVerses/VideoTuna/blob/main/docs/finetune_cogvideox.md.
We are currently refactoring the finetuning pipeline for Hunyuan I2V and will update this functionality very soon. In the meantime, you can try finetuning CogVideo5B-I2V first. Let us know if you encounter any problems during training. Thanks!