The verification results are poor, can you give some advice? Thank you

[goalinshi]

Your work is great, I'm interested in your work, but I tested my own video image effect is very poor, and the coherence after generating the image synthesis video is also poor, the model loading (strict=True) is normal, I'm not sure if it's me doing it wrong, or if the model itself has limitations, please advise

[Boese0601]

Could you please post your results here? Thanks. It would be quite difficult to tell where the error is without visualizations.

[goalinshi]

The coherence of the generated video is poor, and the jitter is serious, but it cannot be well displayed to you;

[songzr0]

I have the same question. FaceID in Generated images does not match the person in the original image