Fine-Tuning AnimateDiff

[laetokang]

Hello!

I am looking to fine-tune the AnimateDiff model with the dataset I've collected.
The "name" variable for my video data is composed of multiple words or phrases, rather than a single clear sentence.
Is it okay to train the model without preprocessing this data?

Thank you!