UNet2DConditionModel参数量不同

BadToBest / EchoMimic

Lifelike Audio-Driven Portrait Animations through Editable Landmark Conditioning

https://badtobest.github.io/echomimic.html

Apache License 2.0

2.68k stars 315 forks source link

Closed aidenyzhang closed 2 months ago

aidenyzhang commented 3 months ago

请教一下，unet_2d_condition.py是做了哪些改动吗？发现跟diffuser包里在同一个config下的参数量不一样。

基于diffusers 0.24.0

基于 src.models.unet_2d_condition import UNet2DConditionModel

JoeFannie commented 2 months ago

只用了部分的权重来作为reference unet的初始化。