No positional embeddings in LatteT2V?

Vchitect / Latte

Latte: Latent Diffusion Transformer for Video Generation.

Apache License 2.0

1.44k stars 147 forks source link

No positional embeddings in LatteT2V? #70

Closed DanielSHKao closed 2 months ago

DanielSHKao commented 2 months ago

The paper mentioned two types of temporal positional embeddings. It seems that they are not included in LatteT2V. Do they deteriorate the generation quality?

Thanks in advance.