Open gsc1456 opened 1 month ago
The generated video cannot fully identify the hair, headwear or hat, and will be distorted, with half moving and half still.
Yes, it’s a known issue! One of the possible reasons is that the training data does not include, or has very few, adornments such as hats. @gsc1456
The generated video cannot fully identify the hair, headwear or hat, and will be distorted, with half moving and half still.