Yujie Wei

@Nancissa Hi! We have provided code that calculates metrics for `CLIP-T`, `CLIP-I`, `DINO-I`, and `Temporal Consistency`. Please refer to [metric/README.MD](https://github.com/ali-vilab/VGen/blob/main/metric/README.MD) for more details.

Hello, I could run DreamVideo without errors by following the environment setup in [the Installation guide](https://github.com/ali-vilab/VGen?tab=readme-ov-file#installation). Please check your environment and look for any potential package conflicts. If you only...

Hello, I could run DreamVideo without errors by following the environment setup in [the Installation guide](https://github.com/ali-vilab/VGen?tab=readme-ov-file#installation). Please check your environment and look for any potential package conflicts. If you only...

Hi, a larger value of p_image_zero means a weaker effect of appearance guidance. You can first try the two endpoints 0 or 0.5, and then adjust it gradually for better...

The results appear to show that the subject ID is overly preserved. You might try reducing the `identity_adapter_index` and increasing the `motion_adapter_index` during inference.

Yes, you will need to adjust the hyperparameters to suit your specific cases and we also provide some examples.

Hi, thanks for your interest. Appearance guidance prevents the model from taking shortcuts (in learning appearance/background) by actively providing prior appearance information of motion videos, enabling the model to focus...