This lora is trained on the wan480p i2v model, and the effect may be a little flawed on 720p. This lora can make the character walk out of the standard model step. Of course, the requirement for the picture is to take a full-body photo as much as possible, standing posture, of course, half-body will also have a good effect, but the action may not be so standard. The trigger word is: catwalk, A woman walked like a model, the weight is recommended to be 0.85-1, and the effect on real portraits is better than that on animated portraits.
Description
FAQ
Comments (2)
There is no need to train on the I2V model. If you train on the T2V model, the resulting lora will be able to do T2V and I2V.
如何才能让生成的视频如此清晰?