Litong.G
Litong.G
Did CausalVideoVAE use 17 frames for training, or was it trained with a variable number of frames? I'm asking because I noticed the same model can perform inference with any...
有没有生成视频的示例呢？我在项目readme里面并没有看到任何视频训练结果的case
> currently we're using sigmoid sampling for timesteps which seems fine but no one has really ablated whether it leaves fine details out Actually, sigmoid and lognorm are mathematically equivalent....
Well, any idea to alleviate the ghost artifacts? > Temporal compression by 8x can result in significant ghosting artifacts, which cannot be reflected in the evaluation metrics.
> 教程还没写完，预计这一周写完 这部分有什么进展吗？